Highest-quality < $1000 non-DV analog capture card for DVD production?
Dan Harkless
precisely, and let you know what info I've already gleaned on my own.]
Howdy. I'm building a new PC whose primary function (besides the
usual DTP, web browsing, etc.) will be uncompressed /
non-lossy-compressed capture of analog NTSC video for burning to DVD
(i.e., 720x480 interlaced, 60 * 1000/1001 fields per second).
Video sources will be TV programs from my Sony SAT-T60 DirecTiVo, my
VHS video collection, and, sometime in the future, my LD collection
(so I'll need composite and S-video in, but not component).
My primary concern is quality -- for instance, I'll be buying a
videophile or pro [S-]VHS deck with TBC so I can squeeze the maximum
quality out of my VHS tapes and then be confident enough to throw away
the originals. I want to avoid an analog -> DV compression -> MPEG
compression signal chain to avoid quality loss due to transcoding /
DCT artifacts. Likewise, I want to avoid hardware MPEG capture, since
it doesn't allow editing without recompression, and since it's of
lower quality than software MPEG encoding (e.g. TMPGenc) in my price
range. Speaking of which, I'd like to spend less than $1,000 on the
capture hardware/software if possible.
I've been doing a fair amount of research, and it appears there's not
that much available in the way of prosumer analog capture. There's
the low-end stuff like the Hauppage WinTV cards, but being so cheap, I
have to imagine the quality achievable with these is less than
perfect. On the other end of the spectrum there's the high-end analog
cards like the Video Toaster NT and Digital Voodoo (Mac-only, I
guess), but these exceed my budget.
There are analog capture solutions that fall between that range, like
the Canopus DVStorm, Matrox RT2500, and Pinnacle Pro-ONE (bugs me that
they stole this name from the old Sequential Circuits analog synth,
BTW ;^>), but it's my understanding that these cards always encode
analog to DV internally (I guess to keep bandwidth down in their
real-time architectures).
Is it possible to do true uncompressed (or non-lossy-compressed, e.g.
Huffyuv) capture with any of the pro/prosumer capture cards? I
understand that if you ask for uncompressed video with some cards the
internal signal chain is actually analog -> DV -> uncompressed, which
again is unacceptable to me. Perhaps it's possible to use a
third-party capture program like iuVCR (VirtualDub seems to be a no-go
since it doesn't work well with Win2K's WDM/DirectShow drivers) with
these cards to do true analog capture? The only card I know that's
_not_ a possibility for is the Pro-ONE -- to quote
<http://www.pinnaclesys.com/SupportFiles/Pro_ONE_CookBook_10.PDF>:
Pinnacle Pro-ONE features a Pinnacle proprietary software
interface called RAL/HAL. Applications that just support standard
interfaces like Video for Windows or Direct Show will not work
with Pinnacle Pro-ONE if they try to capture, do real-time effects
processing or print to tape. There is no SDK available to support
RAL/HAL within other applications.
So, assuming the above cards do not meet my requirements, the only
pro/prosumer analog capture cards I know of are ViewCast's Osprey line
and Winnov's Videum line.
I understand the Ospreys use the same Bt8x8 chipset as the cheap
Hauppage cards, but that the other components on the cards are of
higher quality, though I haven't ever heard any specifics, and I read
on one site that the Osprey-100 is actually just a Hauppage card with
an "Osprey" sticker on it. I'd probably be most interested in the
Osprey-200, since it has on-board audio capture, which theoretically
should solve the common sync issues when capturing audio using a sound
card. However, I read that the Creative SBLive doesn't like having
another audio capture device present on the system -- not sure if this
also applies to the Creative Audigy Platinum eX I'll be putting in my
system.
As for the Winnov cards, apparently they use a different chipset,
though I'm not sure if it's supposed to be higher-quality than the
Bt8x8 or not. I do know that the cards don't support 720x480 out of
the box, and you have to tweak the registry to get them to do this,
which worries me. I've also heard the Win2K drivers aren't very good.
So, of the cards I've mentioned above, can anyone verify ones that
will suit my requirements, and give any impressions as to which
solution will yield the highest quality? It's a pity if the realtime
analog/DV cards are indeed not an option, since the software bundles
they come with seem to have a very high bang-for-the-buck -- I could
definitely make good use of the high-end NLE and limited-version DVD
authoring and streaming video output software. Especially attractive
is the Pro-ONE, due to Pinnacle's offer of a free copy of Commotion
4.0 if you buy before the end of this month. Beyond my need for
simple analog capture, trimming, and DVD output software, I'd love to
play with real production software in conjunction with my Sony DCR-PC5
MiniDV camera.
And of course if there are cards I haven't mentioned above that would
suit my purposes, please let me know. I've ruled out the ATI
All-In-Wonder because I was unsatisfied with its DualHead support
under Win2K -- unlike the Matrox cards, it doesn't support independent
displays / resolutions / refresh rates. I've also ruled out the HDTV
tuner cards since they all seem to support analog and digital
tuning/display, but recording of digital only (though perhaps this is
solvable with third-party capture software?).
In case it matters, the other salient components of the system I'm
currently planning on are:
CPU: dual Athlon MP 1900+ or 2000+ (depending on mobo)
DVD writable drive: not sure yet, due to the DVD-R[W] vs. DVD+R[W]
format war
HDs: 6 Seagate Barracuda ATA IV 80 GB drives, 2 in RAID 1 config.
for the system drive and 4 in RAID 0 config. for the A/V drive
HD controllers: dual 3Ware Escalade 74x0s, or dual Promise FastTrak
100 TXxs (depending on mobo)
Motherboard: Tyan S2460 or S2466
Soundcard: Creative Audigy Platinum eX sound card (already bought,
during an aborted attempt to build this PC using a VIA chipset
mobo -- never could get it stable)
Video card: Matrox G550 (also already bought)
So, sorry again for the long post, and thanks VERY MUCH in advance for
any input.
--
Dan Harkless
use...@harkless.org
Duncan
to the requirement to recompress for editing.
Have you considered MJpeg? I've used a Matrox Marvel G400, which
has a hardware-MJPG compressor, and the quality is extremely
good.
Of course, MJPG lends itself to non-recompressed editing much
better than inter-frame encoding systems such as MPeg1,2,4 etc.
(as each frame is essentially an I-frame).
What is your target playback system?
Are you going to make consumer-compatible DVDs?
Or does the codec not matter, and you're going to watch
the movies on your monitor?
Tim Newton
> to the requirement to recompress for editing.
>
> Have you considered MJpeg? I've used a Matrox Marvel G400, which
> has a hardware-MJPG compressor, and the quality is extremely
> good.
He already said that he did not want to use DV because the loss in
quality, due to the JPEG compression, was unacceptable. Therefore,
MJPEG hardware compressors are unacceptable too.
BTW, I have the Osprey-220 card. Don't assume it gives you perfect
audio/video sync. It does not. At least in VirtualDub, it doesn't.
But that's probably VirtualDub's fault. However, you can't capture
beyond 2 gbytes with Osprey's capture software.
Tim
Ole Hansen
cheap, I have to > imagine the quality achievable with these is less than
perfect.
\begin{rave_on_wintv}
Hmm. Not sure about that. I get excellent results with the WinTV and Osprey
200 cards (both BT878 chipset). I think the main reason they are so cheap is
that they don't offer much in terms of "features" such as hardware
compression, audio capture, gain control etc. But the "low quality"
reputation is largely undeserved. Especially if your source is VHS. There
are no washed out colors or overexposed captures if these cards are tuned
correctly. Aside from being inexpensive, they are also highly compatible. Of
all available capture cards, these are among the ones with the fewest driver
problems and broadest OS support.
\end{rave_one_wintv}
Anyway. In a way I have been in a similar boat as you are now. I ended up
buying a fairly expensive semi-professional computer, but my video input is
still a simple WinTV. For a while I thought about replacing it because
getting audio in sync without framedrops was very time-consuming. (I ended
up capturing and processing the audio separately.) Then a smart guy modified
VirtualDub so it would synchronize the audio on the fly to the video master
clock - problem solved. Perfect sync forever. As a result, there is no need
to buy an Osprey 200 to get audio sync - which wouldn't be perfect with that
card either. I have run all kinds of test patterns through the WinTV and
could never find a show-stopper - a true quality flaw that would justify
replacing it. One weakness certainly is the composite input. Always use the
S-video input if you can. If your LD does not have a quality comb filter
built in, then that might be a reason not to get a WinTV. But then again,
you could run your LD through that prosumer SVHS VCR you are planning to buy
and let it do the Y/C separation. Don't think just because the BT cards are
inexpensive they are bad. The money you pay for more expensive cards goes
largely towards convenience features. And, as you observe correctly, the
current cards are all geared towards DV, with analog inputs added on as a
"legacy feature". The good analog capture cards have all but disappeared.
Perhaps the best place to look for one is eBay. Then again, whatever you buy
there is likely unsupported on the latest Microsoft platform, and no new
drivers will ever appear because all these nice cards have very closed
specs. Oh well.
If you do feel you need something "better" than a WinTV, get something from
Canopus. With money to spare, I would get a DVStorm with the add-on hardware
MPEG-2 encoder. DV is not as bad a source for subsequent MPEG-2 compression
as one might think, although not quite as clean as Huffyuv/uncompressed.
With sufficient MPEG-2 bitrate, you would probably not notice a difference.
Problems would mainly appear with recompression to low bitrates like VCD or
Divx.
Ole
Dan Harkless
> You say that you don't want to use a hardware-compressor, due
> to the requirement to recompress for editing.
No, I said I don't want a hardware _MPEG2_ compressor, due to the
requirement to recompress for editing, and due to the lower quality of
real-time MPEG compression compared to (potentially multipass)
software compression. Otherwise a board with MPEG2 compression would
be fine, as that's my target codec (everything is going to be burned
to DVD).
> Have you considered MJpeg? I've used a Matrox Marvel G400, which
> has a hardware-MJPG compressor, and the quality is extremely
> good.
>
> Of course, MJPG lends itself to non-recompressed editing much
> better than inter-frame encoding systems such as MPeg1,2,4 etc.
> (as each frame is essentially an I-frame).
MJPEG is comparable to DV, which I said I want to avoid so as not to
get quality loss due to DCT artifacts / transcoding. I'm only
interested in uncompressed / non-lossy compression.
> What is your target playback system?
> Are you going to make consumer-compatible DVDs?
> Or does the codec not matter, and you're going to watch
> the movies on your monitor?
No, I want to burn DVD-Video discs.
--
Dan Harkless
use...@harkless.org
Dan Harkless
> BTW, I have the Osprey-220 card. Don't assume it gives you perfect
> audio/video sync. It does not. At least in VirtualDub, it doesn't.
> But that's probably VirtualDub's fault. However, you can't capture
> beyond 2 gbytes with Osprey's capture software.
Ah. Yeah, that would be a problem. For instance, I'll occasionally
want to burn DVDs of movies played on TV without commercial
interruptions, so 2 or 4 GB limitations are definitely unacceptable.
I'll be formatting my capture array as NTFS to avoid such limitations,
so ideally I'd like to find a capture solution that will take
advantage of NTFS' abilities rather than using the kludgy and somewhat
problematic file-to-file linking method.
BTW, what OS are you using your Osprey-220 with? Have you tried
another third-party capture program, like iuVCR
(<http://www.iulab.com/>)? What sound card are you using, and does it
mind the fact that there's another wave capture device on your system?
Have you ever tried capturing the video with your Osprey and the
audio with your sound card? Did that give you better or worse sync?
Thanks for your input!
--
Dan Harkless
use...@harkless.org
Steve Maki
>He already said that he did not want to use DV because the loss in
>quality, due to the JPEG compression, was unacceptable. Therefore,
>MJPEG hardware compressors are unacceptable too.
>
>BTW, I have the Osprey-220 card. Don't assume it gives you perfect
>audio/video sync. It does not. At least in VirtualDub, it doesn't.
>But that's probably VirtualDub's fault. However, you can't capture
>beyond 2 gbytes with Osprey's capture software.
I've got the Osprey 200 and have captured, as an experiment, an hour
of video with no dropped frames and no loss of A/V sync using Virtual
Dub. As a sidenote, if I use VV3's capture utility, I get one dropped
frame every 10 minutes and 50 seconds for some strange reason.
I might mention that I can see no quality penalty by capturing and
editing DV instead of uncompressed, before mpeg2 encoding. I'm not a
professional, but I've looked pretty hard for problems and I can't
see any.
--
Steve Maki
-----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
Check out our new Unlimited Server. No Download or Time Limits!
-----== Over 80,000 Newsgroups - 19 Different Servers! ==-----
Mike T
collection, 500+ LD collection. I first installed a Matrox G450 dual head
card (no capture). It was pretty easy to setup and had very good TV output,
multiburst test signal only about 1 db (10%) down at 5 MHz, no artifacts.
After much reading on the www I acquired a Sony Media Converter, analog to
DV to Firewire. I have run many video test signals and I have to say that
it works pretty good. I have captured high quality C-band satellite video
signals with it and The artifacts are very low when playing back through the
Sony converter. When I played the captured DV through the Microsoft DV
codec and out through the high quality Matrox G450 TV output, the picture
looked soft, I then measured the bandwidth in this mode and there was much
rolloff of high frequencies, about 60% loss at 5MHz on multiburst. This can
only be attributed to the software codec, as all hardware had been tested
independently. My visual judgement based on 40 years of video
design-evaluation is the DV format is very good for most video sources that
a home user would encounter.
I then bought an ATI AIW Radeon 8500DV to serve as a capture card, TV output
card and RF remote control of computer (nice). My early tests are:
The capture function is excellent, video bandwidth is about as good as the
Sony using multiburst test signal.
The RF remote control is good and does control the computer quite well for a
hand held mouse.
The TV output performance STINKS, I can't believe how bad, I still think I
must have something in there difficult software not set correctly and am
still trying to improve this function as it is the mail failing for this
card ($300).
I have not done much capturing yet, but I did run some short captures using
the ATI software. I set the software for a 720X480 uncompressed capture,
and I got about 20% dropped frames using a 7200 rpm ATA100 drive, with 47 GB
of space on the drive the software reported that I could capture about 35
minutes of video. I have added a 160GB raid stripe for more testing. The
capture quality was excellent when I switched back to the Matrox G450 for TV
output. So, you can capture uncompressed if your computer is up to the
daunting task. I will probably use HuffYUV as it is lossless compression
for sources that are good enough.
Your sources of video are not high enough quality to benefit from 720X480
lossless compression. If the Sony SAT-T60 is a DirectTV receiver, then the
compression artifacts and soft picture of directTV will be the limiting
factor, I have a DirectTV system here also and it no where near the quality
of my 12 foot C-band analog dish.
A very good VHS tape is about on par with a very good encoded VCD, I think I
prefer the VCD for its less video ringing and better color bandwidth.
The SVCD's I have encoded look as or better than any of my 500+ Laser Disks
dating back to the first day they went on sale in Atlanta, My player however
is much better than that first Magnavox player :-)
So while it is OK to capture at very high quality, just because you can, it
will cause extra time and use much more of you computer resources than would
"really" be needed. If you plan on converting very many video's to DVD this
will become painfully obvious.
DVD will be my final choice also, but I am waiting out the DVD standards
battle, having already participated in the Beta I-Beta
II-VHS-SuperBeta-SVHS-EdBeta contest.
Gues this is more than enough for now.......
Mike T 40 years of Video design, retired-sorta
--
Please reply to group, if you must reply direct then-
replace $ with s
"Dan Harkless" <use...@harkless.org> wrote in message
news:4189894b.0204...@posting.google.com...
Dan Harkless
> \begin{rave_on_wintv}
> Hmm. Not sure about that. I get excellent results with the WinTV and Osprey
> 200 cards (both BT878 chipset). I think the main reason they are so cheap is
> that they don't offer much in terms of "features" such as hardware
> compression, audio capture, gain control etc. But the "low quality"
> reputation is largely undeserved. Especially if your source is VHS. There
> are no washed out colors or overexposed captures if these cards are tuned
> correctly.
What's involved in tuning them correctly?
> Aside from being inexpensive, they are also highly compatible. Of
> all available capture cards, these are among the ones with the fewest driver
> problems and broadest OS support.
> \end{rave_one_wintv}
Hmm. I've read a bunch of online discussions of various and sundry
problems with the drivers (usually problems with the WDM drivers that
are said not to occur with the VFW ones). But that doesn't mean they
aren't "among the ones with the fewest driver problems" -- everything
else may well be worse... Also, I'd imagine either Hauppage or
ViewCast has better drivers than the other company -- anyone have any
views on that?
> Anyway. In a way I have been in a similar boat as you are now. I ended up
> buying a fairly expensive semi-professional computer, but my video input is
> still a simple WinTV. For a while I thought about replacing it because
> getting audio in sync without framedrops was very time-consuming. (I ended
> up capturing and processing the audio separately.)
By "separately" do you mean in a second pass, or...?
> Then a smart guy modified
> VirtualDub so it would synchronize the audio on the fly to the video master
> clock - problem solved. Perfect sync forever.
When you say "a smart guy modified VirtualDub", do you mean there's a
modified version of VirtualDub that someone else distributes, or that
this ability is now folded into the main version? I assume you're
using the VFW drivers for the WinTV (as VirtualDub is said not to work
well with the WDM->VFW wrapper)?
> I have run all kinds of test patterns through the WinTV and
> could never find a show-stopper - a true quality flaw that would justify
> replacing it.
Ah -- that's good to know. Hadn't seen reports elsewhere of
scientific testing of capture cards like this, using test patterns.
> The good analog capture cards have all but disappeared.
> Perhaps the best place to look for one is eBay. Then again, whatever you buy
> there is likely unsupported on the latest Microsoft platform, and no new
> drivers will ever appear because all these nice cards have very closed
> specs. Oh well.
Interesting. What were some of the old classics?
> If you do feel you need something "better" than a WinTV, get something from
> Canopus. With money to spare, I would get a DVStorm with the add-on hardware
> MPEG-2 encoder. DV is not as bad a source for subsequent MPEG-2 compression
> as one might think, although not quite as clean as Huffyuv/uncompressed.
> With sufficient MPEG-2 bitrate, you would probably not notice a difference.
> Problems would mainly appear with recompression to low bitrates like VCD or
> Divx.
Yeah, I can imagine how if you throw enough bits at the MPEG
compression you could at least faithfully reproduce the DCT artifacts
from the DV encoding without losing detail elsewhere in the picture
because you used those bits on the artifacts. But it would seem that
no matter how many bits you throw at the compression, you're always
going to get two levels of DCT artifacts in the analog -> DV -> MPEG
scenario, and thus lower quality than analog -> uncompressed/non-lossy
-> MPEG. I guess what may save the day is that DV uses a pretty high
bitrate and doesn't usually have easily visible artifacts.
--
Dan Harkless
use...@harkless.org
Dan Harkless
> I have started on the same journey with about the same needs, 3000+ VHS tape
> collection, 500+ LD collection. I first installed a Matrox G450 dual head
> card (no capture). It was pretty easy to setup and had very good TV output,
> multiburst test signal only about 1 db (10%) down at 5 MHz, no artifacts.
Presumably my G550 is of similar quality to the G450.
> After much reading on the www I acquired a Sony Media Converter, analog to
> DV to Firewire.
I would imagine this uses about the same circuitry as my same-era Sony
DCR-PC5 MiniDV camcorder, which also has the ability to tapelessly
convert analog to DV.
> I have run many video test signals and I have to say that
> it works pretty good. I have captured high quality C-band satellite video
> signals with it and The artifacts are very low when playing back through the
> Sony converter. When I played the captured DV through the Microsoft DV
> codec and out through the high quality Matrox G450 TV output, the picture
> looked soft, I then measured the bandwidth in this mode and there was much
> rolloff of high frequencies, about 60% loss at 5MHz on multiburst. This can
> only be attributed to the software codec, as all hardware had been tested
> independently. My visual judgement based on 40 years of video
> design-evaluation is the DV format is very good for most video sources that
> a home user would encounter.
I've heard others mention that they got much higher quality with
non-Microsoft DV codecs -- have you tried that? Personally, though,
I'm surprised there are such noticeable quality variations -- I would
have thought you either correctly decode the DV video or you don't. I
wouldn't expect there to be quality variations other than those
induced by bugs (and perhaps that is indeed what we're talking about
-- this is Microsoft, after all).
> I then bought an ATI AIW Radeon 8500DV to serve as a capture card, TV output
> card and RF remote control of computer (nice). My early tests are:
>
> The capture function is excellent, video bandwidth is about as good as the
> Sony using multiburst test signal.
Which one would you give the edge to?
> The RF remote control is good and does control the computer quite well for a
> hand held mouse.
>
> The TV output performance STINKS, I can't believe how bad, I still think I
> must have something in there difficult software not set correctly and am
> still trying to improve this function as it is the mail failing for this
> card ($300).
>
> I have not done much capturing yet, but I did run some short captures using
> the ATI software. I set the software for a 720X480 uncompressed capture,
> and I got about 20% dropped frames using a 7200 rpm ATA100 drive, with 47 GB
> of space on the drive the software reported that I could capture about 35
> minutes of video. I have added a 160GB raid stripe for more testing. The
> capture quality was excellent when I switched back to the Matrox G450 for TV
> output. So, you can capture uncompressed if your computer is up to the
> daunting task. I will probably use HuffYUV as it is lossless compression
> for sources that are good enough.
>
> Your sources of video are not high enough quality to benefit from 720X480
> lossless compression. If the Sony SAT-T60 is a DirectTV receiver,
It is.
> then the
> compression artifacts and soft picture of directTV will be the limiting
> factor, I have a DirectTV system here also and it no where near the quality
> of my 12 foot C-band analog dish.
Okay. I guess I was just concerned about having three layers of DCT
artifacts (DirecTV MPEG -> DV -> DVD MPEG). But then I suppose the
blocking artifacts of the DirecTV MPEG will actually *help* the
subsequent compression, and it's only the mosquito noise artifacts I'd
have to be worried about.
> A very good VHS tape is about on par with a very good encoded VCD, I think I
> prefer the VCD for its less video ringing and better color bandwidth.
>
> The SVCD's I have encoded look as or better than any of my 500+ Laser Disks
> dating back to the first day they went on sale in Atlanta, My player however
> is much better than that first Magnavox player :-)
>
> So while it is OK to capture at very high quality, just because you can, it
> will cause extra time and use much more of you computer resources than would
> "really" be needed. If you plan on converting very many video's to DVD this
> will become painfully obvious.
Well, my library is nowhere near as large as yours -- several hundred
rather than several thousand. As I mentioned, part of the reason I
wanted to "overspec" the capture was so I could confidently throw out
the original tapes after digitization.
> DVD will be my final choice also, but I am waiting out the DVD standards
> battle, having already participated in the Beta I-Beta
> II-VHS-SuperBeta-SVHS-EdBeta contest.
I'd like to get started soon, so I'm going to have to make a leap of
faith -- I'm leaning towards DVD+R[W], but I wish there was more
information available about whether the newly-released DVD+R discs
rival the DVD-R ones in player compatibility.
> Gues this is more than enough for now.......
>
> Mike T 40 years of Video design, retired-sorta
Thanks a lot for the input, Mike! Your comments and others' have made
me reconsider the viability of an analog -> DV -> MPEG signal chain.
Certainly no one has chimed in and said, "Yeah, I've compared
uncompressed analog -> MPEG vs. analog -> DV -> MPEG and gotten
noticeably higher-quality results with the former" (at least in this
thread -- I have seen comments to that effect on rec.video.desktop in
the past).
I guess one flaw in my original thinking is not acknowledging that if
DV(25) -> MPEG is good enough for some recent independent "film"
productions, it should be good enough for my analog video library
digitization.
--
Dan Harkless
use...@harkless.org
Ole Hansen
> What's involved in tuning them correctly?
Mainly adjusting brightness and contrast correctly, maybe also color
saturation. VirtualDub has a mode called "Preview with histogram" that
helps with this. It is the equivalent of setting the record level on a
cassette deck.
> Hmm. I've read a bunch of online discussions of various and sundry
> problems with the drivers (usually problems with the WDM drivers that
> are said not to occur with the VFW ones). But that doesn't mean they
> aren't "among the ones with the fewest driver problems" -- everything
> else may well be worse...
That's what I mean. Everything else appears to be worse. BTW, I use the
WinTV with the open source WDM drivers from
http://btwincap.sourceforge.net
without any real problems (just minor convenience issues).
> Also, I'd imagine either Hauppage or
> ViewCast has better drivers than the other company -- anyone have any
> views on that?
I recently tested an Osprey 200 and found Viewcast's drivers much better
than Hauppauge's for the WinTV. Interestingly, the Win2k driver for the
WinTV uses what appears to be an older version of the Osprey 100 drivers
...
> (I ended up capturing and processing the audio separately.)
>
> By "separately" do you mean in a second pass, or...?
Actually, I captured the audio simultaneously to a separate disk. It
worked, amazingly, without interfering with the video capture.
> When you say "a smart guy modified VirtualDub", do you mean there's a
> modified version of VirtualDub that someone else distributes,
Yes. It's here:
http://www-user.rhrk.uni-kl.de/~dittrich/sync/
> or that this ability is now folded into the main version?
It's not (yet).
> I assume you're using the VFW drivers for the WinTV (as VirtualDub is said not to work
> well with the WDM->VFW wrapper)?
No, I'm using the open source drivers mentioned above, which use the WDM
architecture. VirtualDub seems to have no problems with the wrapper under
Win2k.
> Interesting. What were some of the old classics?
Some that come to mind:
Pinnacle/Miro DC30, DC50. Fast AV Master, AV Master 2000.
Matrox Rainbow Runner Studio, Rainbow Runner G series, Marvel G200/G400.
These are all MJPEG cards, but some (all?) of them can also deliver
uncompressed (YUY2, RGB24), which is what you want. I know the Matrox
cards can.
> Yeah, I can imagine how if you throw enough bits at the MPEG
> compression you could at least faithfully reproduce the DCT artifacts
> from the DV encoding without losing detail elsewhere in the picture
> because you used those bits on the artifacts. But it would seem that
> no matter how many bits you throw at the compression, you're always
> going to get two levels of DCT artifacts in the analog -> DV -> MPEG
> scenario, and thus lower quality than analog -> uncompressed/non-lossy
> -> MPEG. I guess what may save the day is that DV uses a pretty high
> bitrate and doesn't usually have easily visible artifacts.
I agree. Same thinking here. Although I haven't tried, I bet the route via
DV will be good enough for VHS sources. Whatever DV clips I have are
practically free of DCT artifacts, even when enlarged. It is surely also a
function of the DV codec. I hear that the one by Canopus is among the
best.
Ole
Mike T
Mike T
--
Please reply to group, if you must reply direct then-
replace $ with s
"Dan Harkless" <use...@harkless.org> wrote in message
news:4189894b.02041...@posting.google.com...
> "Mike T" <mtallent@high$tream.net> wrote:
> > I have started on the same journey with about the same needs, 3000+ VHS
tape
> > collection, 500+ LD collection. I first installed a Matrox G450 dual
head
> > card (no capture). It was pretty easy to setup and had very good TV
output,
> > multiburst test signal only about 1 db (10%) down at 5 MHz, no
artifacts.
>
> Presumably my G550 is of similar quality to the G450.
Probably, I don't know for sure.
>
> > After much reading on the www I acquired a Sony Media Converter, analog
to
> > DV to Firewire.
>
> I would imagine this uses about the same circuitry as my same-era Sony
> DCR-PC5 MiniDV camcorder, which also has the ability to tapelessly
> convert analog to DV.
I would suggest you try some short captures of your different video sources
with the Sony DV cam and see how they look, you will probobly not be able to
see much difference. I think the Sony Media converter is no longer
available from Sony.
I was really surprised at the difference with the software codec, I have the
Main Concept and will be testing it also. The problem I think, is most
don't seem to test these things as video devices using actual video test
signals. I have found very few video specifications in any of this computer
hardware or software.
>
> > I then bought an ATI AIW Radeon 8500DV to serve as a capture card, TV
output
> > card and RF remote control of computer (nice). My early tests are:
> >
> > The capture function is excellent, video bandwidth is about as good as
the
> > Sony using multiburst test signal.
>
> Which one would you give the edge to?
Aa for video bandwidth, I think they are very close, I plan to take pictures
of waveforms when I get everything together to test.
The ATI has the advantage of being able to select many different types of
captures and pixel rates, but it does not include sound and I have not done
long captures yet to check for sound sync problems. This might be the big
problem with video capture only devices. The Sony converter does video and
PCM sound together.
I used the Sony to capture the 911 show on CBS and I planned to use the
C-band analog feed, but it was windy and I switched to the DirectTV dish.
Captured 27 GB to computer and then converted it to xVCD format with
TMPEGenc on 2 disks. I have to say it looks pretty good, better than VHS,
the mpeg artifacts from DirectTv are still there, but I don't think they are
much worse than the original looked on my TV monitor. The detail is a
little softer than original due to the lower sample rate of VCD format, but
a little better than VHS but no tape drop-outs and good color bandwidth.
Since I was only putting 60 minutes on a 80 minute CD, I increased the
bitrate to use all the available space on the CDR, this helps the VCD
quality. Also TMPEGenc has some block noise filters that help a little.
>
> > A very good VHS tape is about on par with a very good encoded VCD, I
think I
> > prefer the VCD for its less video ringing and better color bandwidth.
> >
> > The SVCD's I have encoded look as or better than any of my 500+ Laser
Disks
> > dating back to the first day they went on sale in Atlanta, My player
however
> > is much better than that first Magnavox player :-)
> >
> > So while it is OK to capture at very high quality, just because you can,
it
> > will cause extra time and use much more of you computer resources than
would
> > "really" be needed. If you plan on converting very many video's to DVD
this
> > will become painfully obvious.
>
> Well, my library is nowhere near as large as yours -- several hundred
> rather than several thousand. As I mentioned, part of the reason I
> wanted to "overspec" the capture was so I could confidently throw out
> the original tapes after digitization.
I have been recording mostly movies on VHS for over 20 years, so I will
probably only convert the more obscure ones not yet available on DVD.
Fortunantly I have weened myself from recording very much lately
>
> > DVD will be my final choice also, but I am waiting out the DVD standards
> > battle, having already participated in the Beta I-Beta
> > II-VHS-SuperBeta-SVHS-EdBeta contest.
>
> I'd like to get started soon, so I'm going to have to make a leap of
> faith -- I'm leaning towards DVD+R[W], but I wish there was more
> information available about whether the newly-released DVD+R discs
> rival the DVD-R ones in player compatibility.
I've read a lot of conflicting statements with regard to compatability about
the -R and +R disks, that's why I decided to wait awhile.
Dan Harkless
> That's what I mean. Everything else appears to be worse. BTW, I use the
> WinTV with the open source WDM drivers from
>
> http://btwincap.sourceforge.net
Ah, thanks very much for the link -- I hadn't come across any
reference to this third-party driver (just the less featureful
iulab.com one). Pity it doesn't support the onboard audio capture of
Bt87x cards like the Osprey 2x0, though. But I suppose the audio ADCs
on those cards may or may not be as good as the ones on the Creative
Labs Audigy and other high-quality sound cards.
> I recently tested an Osprey 200 and found Viewcast's drivers much better
> than Hauppauge's for the WinTV.
Good to know -- any specifics you can mention?
> Interestingly, the Win2k driver for the
> WinTV uses what appears to be an older version of the Osprey 100 drivers
Not too surprising, since, as I noted in another post, the Osprey 100
board reportedly has a "Hauppage" logo under the "Osprey" sticker.
> > When you say "a smart guy modified VirtualDub", do you mean there's a
> > modified version of VirtualDub that someone else distributes,
>
> Yes. It's here:
>
> http://www-user.rhrk.uni-kl.de/~dittrich/sync/
Thanks a lot for this link as well! Another one I hadn't come across
in my research. The audio sample rate conversion done by VirtualDub
Sync makes a lot more sense than the duplicated and/or dropped frames
used by the original VirtualDub (and "all capture programs we know",
according to the VirtualDub Sync guy). Reminds me of the "Night of
the Living Dead" DVD put out by Elite (the old one -- not sure whether
the new "Millenium" edition fixes this) -- to keep audio and video
(from different elements) in sync, they elected to duplicate frames,
which results in a distracting periodic freeze of the video. Would
have made much more sense to mangle the audio track to fit the video,
rather than vice-versa.
> > or that this ability is now folded into the main version?
>
> It's not (yet).
Surprising. Hopefully there's interest from Avery Lee in doing this.
> > I assume you're using the VFW drivers for the WinTV (as VirtualDub is said
> > not to work well with the WDM->VFW wrapper)?
>
> No, I'm using the open source drivers mentioned above, which use the WDM
> architecture. VirtualDub seems to have no problems with the wrapper under
> Win2k.
Excellent!
> > Interesting. What were some of the old classics?
>
> Some that come to mind:
> Pinnacle/Miro DC30, DC50. Fast AV Master, AV Master 2000.
> Matrox Rainbow Runner Studio, Rainbow Runner G series, Marvel G200/G400.
>
> These are all MJPEG cards, but some (all?) of them can also deliver
> uncompressed (YUY2, RGB24), which is what you want. I know the Matrox
> cards can.
Okay. I've heard of a few of those. So did you mean that some or all
of those had significantly better raw video quality than what's
available today?
> I agree. Same thinking here. Although I haven't tried, I bet the route via
> DV will be good enough for VHS sources. Whatever DV clips I have are
> practically free of DCT artifacts, even when enlarged. It is surely also a
> function of the DV codec. I hear that the one by Canopus is among the
> best.
Good to know. I'll consider that if I elect to go the RT DV route.
Of course the offer of a free copy of Commotion with the Pinnacle
Pro-ONE is likely to sway me more than the better DV codec and other
few advantages of the Canopus DVStorm.
--
Dan Harkless
use...@harkless.org
Dan Harkless
> I would suggest you try some short captures of your different video sources
> with the Sony DV cam and see how they look, you will probobly not be able to
> see much difference.
Not having the DVD mastering solution in place yet, I've been
archiving some DirecTV broadcasts to MiniDV tape, and I certainly
haven't noticed any quality problems (other than audio dropouts I was
getting for awhile due to a tape mechanism problem I got repaired),
though I haven't any real critical viewing, nor used test patterns or
anything.
> The ATI has the advantage of being able to select many different types of
> captures and pixel rates, but it does not include sound and I have not done
> long captures yet to check for sound sync problems. This might be the big
> problem with video capture only devices. The Sony converter does video and
> PCM sound together.
Yes, the interleaved audio and video is definitely a nice boon of the
DV format. Certainly a signal with perfect A/V sync but a few extra
DCT artifacts is preferable to a cleaner video signal with drifting
audio.
> Also TMPEGenc has some block noise filters that help a little.
Interesting. I can't quite imagine how you could "filter" block noise
once it's occurred. Seems like there'd be no way to distinguish
artifactual (is that a word?) blocks from intentional blocks of color.
--
Dan Harkless
use...@harkless.org
MHz-man
> you could run your LD through that prosumer SVHS VCR you are planning to buy
> and let it do the Y/C separation.
This highers the price a whole lot.
What do you think about this:
http://www.smarthome.com/8290.html
or this:
http://www.smarthome.com/8299.html
?
I'm wondering if those devices can work in both pal and ntsc or only on one of the two standards...
Ole Hansen
> > than Hauppauge's for the WinTV.
>
> Good to know -- any specifics you can mention?
First, I guess my statement sounds a bit misleading. I didn't try the
drivers for the Osprey 200 with the WinTV. What I am comparing is an Osprey
200 with the drivers from Viewcast that are included in the package to a
WinTV with the WinTV drivers from Hauppauge's Web site (latest drivers for
Win2K: W2KDRV311.EXE). Both are VfW drivers. I do not know if one can
operate a WinTV with Osprey drivers - if so, then most likely without tuner
support.
Specific differences: Hauppauge drivers occasionally drop fields/reverse
field order (even under Win98). That does not happen with the Osprey 200.
(The WinTV does work ok (no dropped fields) with the open-source WDM driver,
btwincap). In addition, the dialog windows for the Osprey 200 are much
clearer and provide more options, in particular a setting for CCIR601
(non-square pixel) sampling rate, and custom frame sizes. These settings
turn out to be accessible with the Hauppauge drivers, too, but they are
deeply hidden. Also, the Osprey drivers come pre-tuned - contrast/brightness
appear to have been calibrated at the factory. The driver does not use the
defaults ("127"), but different values (Brightness 133, Contrast 110 for
S-video, slightly lower for composite). Finally, the Osprey does not give
any black bars on either side of the picture, unlike the WinTV. It seems to
observe the selected aspect ratio correctly. All in all, the Osprey driver
gives the impression that it was put together with more care, and it works
better (in particular, no field order problems).
> > Interestingly, the Win2k driver for the
> > WinTV uses what appears to be an older version of the Osprey 100 drivers
>
> Not too surprising, since, as I noted in another post, the Osprey 100
> board reportedly has a "Hauppage" logo under the "Osprey" sticker.
Indeed, my Osprey 200 has a Hauppauge logo under a silly white sticker, too.
> > Some that come to mind:
> > Pinnacle/Miro DC30, DC50. Fast AV Master, AV Master 2000.
> > Matrox Rainbow Runner Studio, Rainbow Runner G series, Marvel G200/G400.
> >
> > These are all MJPEG cards, but some (all?) of them can also deliver
> > uncompressed (YUY2, RGB24), which is what you want. I know the Matrox
> > cards can.
>
> Okay. I've heard of a few of those. So did you mean that some or all
> of those had significantly better raw video quality than what's
> available today?
No. The raw digitizers are probably similar in quality to today's analog->DV
hardware, and MJPEG is about as good as DV in terms of quality. (I
understand DV is essentially an improved form of MJPEG.) However, some of
these cards also provide raw (uncompressed) data, and that may have higher
quality (fewer compression artifacts) than MJPEG/DV. I do not know, in fact
I doubt, that today's DV cards can provide true uncompressed output (i.e.
data that were not DV-compressed first).
> As for the Winnov cards, apparently they use a different chipset,
> though I'm not sure if it's supposed to be higher-quality than the
> Bt8x8 or not. I do know that the cards don't support 720x480 out of
> the box, and you have to tweak the registry to get them to do this,
> which worries me. I've also heard the Win2K drivers aren't very good.
I had a bad experience with a Winnov Videum AV. 640x480 was actually
640x240, interpolated to 480 lines. 30fps was not possible without dropping
frames, no matter what. And the native capture format appeared to be their
proprietary WNV1 hardware-compressed format. If you asked for YUY2, you got
analog->WNV1->YUY2. WNV1 was termed "lossless", but actually involved
reduction of the color depth to 7 bits/color. Unacceptable to me. Winnov
have come out with a second-generation board (Videum 1000, IIRC) that
probably does not have some of these problems, but the information on the
Web is insufficient to conclude that for sure. Winnov also appear to have a
tendency for "creative marketing" - IOW, read between the lines.
Cheers,
Ole
Ole Hansen
news:a9rkg...@enews4.newsguy.com...
> > But then again,
> > you could run your LD through that prosumer SVHS VCR you are planning to
buy
> > and let it do the Y/C separation.
>
> This highers the price a whole lot.
>
No, since he is planning to get the deck anyway.
> What do you think about this:
> http://www.smarthome.com/8290.html
>
> or this:
> http://www.smarthome.com/8299.html
>
> ?
Stay away. Far away. For that price, buy a basic JVC SVHS VCR. Same or
better Y/C separation, plus a VCR.
> I'm wondering if those devices can work in both pal and ntsc or only on
one of the two standards...
Will only work with NTSC. Absolutely sure.
Ole
MHz-man
Just...
> > Okay. I've heard of a few of those. So did you mean that some or all
> > of those had significantly better raw video quality than what's
> > available today?
>
> No. The raw digitizers are probably similar in quality to today's analog->DV
> hardware, and MJPEG is about as good as DV in terms of quality. (I
> understand DV is essentially an improved form of MJPEG.) However, some of
> these cards also provide raw (uncompressed) data, and that may have higher
> quality (fewer compression artifacts) than MJPEG/DV. I do not know, in fact
> I doubt, that today's DV cards can provide true uncompressed output (i.e.
> data that were not DV-compressed first).
... I guess that was not exactly the question ... or I didn't understand it completely?
I guess Dan was asking if the raw capture (no compression) of the old classic analog capture cards
was better than the raw capture of the modern analog cards (not the DV ones, the analog ones like
e.g. the Brooktree ones which I'm very interested into).
I was really interested in your reply to this question, could you tell us your opinion?
Thanks!
Ole Hansen
>
> I guess Dan was asking if the raw capture (no compression) of the old
classic
> analog capture cards was better than the raw capture of the modern analog
> cards (not the DV ones, the analog ones like e.g. the Brooktree ones which
I'm
> very interested into).
>
> I was really interested in your reply to this question, could you tell us
your opinion?
The only "classic" card I have direct experience with is the Matrox Rainbow
Runner Studio (a 1997 model). It uses a Philips video decoder (SAA7111A).
The other cards I mentioned probably use similar chips. (Philips markets a
variety of decoders/encoders, which seem to have become some sort of de
facto consumer-level industry standard.) They all yield about the same raw
video quality: 8-bit ADC resolution, 4:2:2 color sampling, YUY2 native
format, notch-type comb filter for composite. Now, the specs for the BT8xx
are essentially the same, except that the BT chips include a few extras like
a PCI interface and audio circuirty, things that are extraneous to the video
signal path. There may be other differences like gain control, sampling
frequencies available, resampling algorithms used, etc. A study of the data
sheets may clarify some of these (or confuse even more). In any case, the
basic specs are going to be all the same for consumer/prosumer-type capture
cards. For instance, no sane engineer would put a 12-bit decoder on a card
designed to output YUY2 or even RGB24 raw video.
So, whatever video quality differences there might be between cards, they
are likely to be minor. The raw captures from my old Rainbow Runner were not
noticably different from what I get out of the WinTV. I did test captures
with the WinTV that demonstrate 500+ lines of resolution with S-video input,
and that it about as good as it gets.
To prove the point, I feel inspired to dig out my old Rainbow Runner and do
a direct comparison. I don't have much time, but when I do this, I'll post
the results.
Also, don't forget the specs of your sources and targets. If you are
thinking VHS -> MPEG-2, for instance, the quality of the capture card's
video decoder is not going to matter much.
Hope this is useful,
Ole
MHz-man
> video quality: 8-bit ADC resolution, 4:2:2 color sampling, YUY2 native
> format,
I have an hardware-broken Matrox Rainbow Runner, which cannot deliver mjpeg (blurred images), cannot
deliver YUY2 (50% of frames are 90% green + other crazy problems), but correctly delivers RGB24.
This leads me to think that the sampling is RGB24 (which is 4:4:4, right?) and then the broken parts
of the card are the mjpeg capture and the YUY2 conversion.
However I can't be sure of this, if you have read the technical specs you are probably right.
> notch-type comb filter for composite.
Excuse the ignorance, maybe I can't understand English language well enough
What is a "notch-type comb filter"? Is it good, better than BT8xx?
> Now, the specs for the BT8xx are essentially the same
> [...]
Thanks A-LOT
So I will buy a BT8xx card
> To prove the point, I feel inspired to dig out my old Rainbow Runner and do
> a direct comparison. I don't have much time, but when I do this, I'll post
> the results.
Arf...! (extra salivation...)
This would be great. But take your time, I will have already bought a Brooktree card by that time
:-)
Even then I will check this newsgroup from time to time for your report!
Just curiosity... Do you think semi-pro analog capture cards have a better sampler than Brooktree?
And what about professional ones? At which level of $$$ there is an improvement on raw data? (and
which improvement might it be, a 4:4:4 sampling maybe?)
Thanks again Ole.
Dan Harkless
> > > I recently tested an Osprey 200 and found Viewcast's drivers much better
> > > than Hauppauge's for the WinTV.
> >
> > Good to know -- any specifics you can mention?
>
> First, I guess my statement sounds a bit misleading. I didn't try the
> drivers for the Osprey 200 with the WinTV. What I am comparing is an Osprey
> 200 with the drivers from Viewcast that are included in the package to a
> WinTV with the WinTV drivers from Hauppauge's Web site (latest drivers for
> Win2K: W2KDRV311.EXE).
Yeah, I understood what you meant.
> Both are VfW drivers. I do not know if one can operate a WinTV with Osprey
> drivers - if so, then most likely without tuner support.
That is an interesting question -- I hadn't thought of that HW/SW
combination. This page for the open source WDM Bt8xx drivers talks
about some of the issues that can occur if you use settings for a
different Bt8xx card than you actually have:
http://btwincap.sourceforge.net/custom.html
Problems would probably be minimized, however, for the cards that
Hauppage OEMs for ViewCast.
> Specific differences: Hauppauge drivers occasionally drop fields/reverse
> field order (even under Win98). That does not happen with the Osprey 200.
> (The WinTV does work ok (no dropped fields) with the open-source WDM driver,
> btwincap). In addition, the dialog windows for the Osprey 200 are much
> clearer and provide more options, in particular a setting for CCIR601
> (non-square pixel) sampling rate, and custom frame sizes. These settings
> turn out to be accessible with the Hauppauge drivers, too, but they are
> deeply hidden. Also, the Osprey drivers come pre-tuned - contrast/brightness
> appear to have been calibrated at the factory. The driver does not use the
> defaults ("127"), but different values (Brightness 133, Contrast 110 for
> S-video, slightly lower for composite). Finally, the Osprey does not give
> any black bars on either side of the picture, unlike the WinTV. It seems to
> observe the selected aspect ratio correctly. All in all, the Osprey driver
> gives the impression that it was put together with more care, and it works
> better (in particular, no field order problems).
Wow! Those are very significant quality differences. I guess I will
exclusively consider the Ospreys for uncompressed capture, if it is
indeed the case that none of the pro DV cards with analog input allow
you to get at the signal prior to it being DV-compressed (seems quite
an unnecessary shame if that's how all those cards are designed).
> > Not too surprising, since, as I noted in another post, the Osprey 100
> > board reportedly has a "Hauppage" logo under the "Osprey" sticker.
>
> Indeed, my Osprey 200 has a Hauppauge logo under a silly white sticker, too.
Very odd. I wonder whether the entire Osprey line is OEM'd by
Hauppage... Given this, I wonder if it's really true, as some have
suggested, that the components on the Ospreys other than the Bt8xx
chip are of higher quality than those on the Hauppages, or whether
Hauppage just uses the same components for both in-house and OEM
production.
But I suppose that's a moot point if the Osprey *drivers* have all the
advantages that you mention. Perhaps ViewCast is more of a software
house than a hardware operation.
BTW, Tim Newton mentioned in his post in this thread that one
shouldn't expect perfect audio sync from the Osprey-2x0, even though
it presumably uses a master oscillator for both audio and video. Is
this your experience as well, or is there a capture program one can
use (e.g. iuVCR) with the board that does the correct 48 kHz 8008
samples per 5 frames of 29.97 NTSC (see DV magazine's great article
"Keeping That Syncing Feeling",
<http://www.dv.com/features/features_item.jhtml?LookupId=/dv/xml/feature/2001/meyer0901>)
and that also breaks the 2 GB and 4 GB filesize limits (unlike the
Osprey's bundled capture software, per Tim)?
> I had a bad experience with a Winnov Videum AV. 640x480 was actually
> 640x240, interpolated to 480 lines. 30fps was not possible without dropping
> frames, no matter what. And the native capture format appeared to be their
> proprietary WNV1 hardware-compressed format. If you asked for YUY2, you got
> analog->WNV1->YUY2.
Ouch! Okay, they're definitely off my list. Thanks for the info!!
--
Dan Harkless
use...@harkless.org
Dan Harkless
> > I guess Dan was asking if the raw capture (no compression) of the old
> > classic analog capture cards was better than the raw capture of the modern
> > analog cards (not the DV ones, the analog ones like e.g. the Brooktree ones
> > which I'm very interested into).
Yeah, that's what I was asking.
> The only "classic" card I have direct experience with is the Matrox Rainbow
> Runner Studio (a 1997 model). It uses a Philips video decoder (SAA7111A).
> The other cards I mentioned probably use similar chips. (Philips markets a
> variety of decoders/encoders, which seem to have become some sort of de
> facto consumer-level industry standard.) They all yield about the same raw
> video quality: 8-bit ADC resolution, 4:2:2 color sampling, YUY2 native
> format, notch-type comb filter for composite. Now, the specs for the BT8xx
> are essentially the same, except that the BT chips include a few extras like
> a PCI interface
Huh? Wouldn't all these cards need a PCI interface?
Oh... I get it -- you mean the PCI interface is on-chip for the
Bt8xxs.
> and audio circuirty, things that are extraneous to the video
> signal path. There may be other differences like gain control, sampling
> frequencies available, resampling algorithms used, etc. A study of the data
> sheets may clarify some of these (or confuse even more). In any case, the
> basic specs are going to be all the same for consumer/prosumer-type capture
> cards. For instance, no sane engineer would put a 12-bit decoder on a card
> designed to output YUY2 or even RGB24 raw video.
>
> So, whatever video quality differences there might be between cards, they
> are likely to be minor.
Short of any comparative testing results that may exist out there (I
haven't come across any), I guess your theorizing will have to be
enough to quell my anal perfectionistic compulsions for the *BEST*
quality solution. ;^>
> The raw captures from my old Rainbow Runner were not
> noticably different from what I get out of the WinTV. I did test captures
> with the WinTV that demonstrate 500+ lines of resolution with S-video input,
> and that it about as good as it gets.
Well, more to capture quality than lines of resolution, of course (I
know you didn't mean to imply otherwise), but yeah, that is a good
sign.
> To prove the point, I feel inspired to dig out my old Rainbow Runner and do
> a direct comparison. I don't have much time, but when I do this, I'll post
> the results.
I hate to see you go out of your way, but if you do feel like doing
that, that'd be very illuminating. Hopefully you can post the results
(or at least a pointer to them) as a followup to this thread so no one
following it misses them?
> Also, don't forget the specs of your sources and targets. If you are
> thinking VHS -> MPEG-2, for instance, the quality of the capture card's
> video decoder is not going to matter much.
My VHS library will make up a big chunk of my digitization activities,
but I also want a solution that will be able to handle the
superior-to-VHS quality of my DirecTiVo and (especially) my
LaserDiscs.
--
Dan Harkless
use...@harkless.org
Dan Harkless
> Excuse the ignorance, maybe I can't understand English language well enough
> What is a "notch-type comb filter"? Is it good, better than BT8xx?
Sounds to me like it means a comb filter that's not a true multi-notch
comb filter, but rather a single notch. This'd be inferior in quality
to a real comb filter, but if it matters to you, just do your own comb
filtering of composite signals and use the S-video input.
> And what about professional ones? At which level of $$$ there is an
> improvement on raw data? (and which improvement might it be, a 4:4:4 sampling
> maybe?)
As documented on the OneRiver Media codec review page:
http://www.onerivermedia.com/codecs/
specifically, in the "Purpose" popup, even the ultra-high-end
"uncompressed" capture boards are in fact compressed 2:1 in the color
space, to 4:2:2. According to the author, there are no 4:4:4 boards
at this time.
I'm not sure which of the uncompressed boards is the cheapest, but I
believe they're all multiple thousands of dollars.
--
Dan Harkless
use...@harkless.org
Dan Harkless
> What do you think about this:
> http://www.smarthome.com/8290.html
>
> or this:
> http://www.smarthome.com/8299.html
I'd have to think any passive composite <-> S-video converter would be
unable to match the potential quality of an active solution.
And I don't really understand the signal theory involved, but I
thought dot crawl was inherent to the composite format, so claims of
jamming an S-video signal into a composite jack without loss of
quality seem very suspect to me.
Certainly with my ~$20 Radio Shack composite <-> S-video converter,
taking a pristine S-video signal and converting it to composite
introduces the usual composite video artifacts.
--
Dan Harkless
use...@harkless.org
John S. Dyson
> "MHz-man" <MHz...@no.spam.com> wrote:
> > What do you think about this:
> > http://www.smarthome.com/8290.html
> >
> > or this:
> > http://www.smarthome.com/8299.html
>
> I'd have to think any passive composite <-> S-video converter would be
> unable to match the potential quality of an active solution.
>
> And I don't really understand the signal theory involved, but I
> thought dot crawl was inherent to the composite format, so claims of
> jamming an S-video signal into a composite jack without loss of
> quality seem very suspect to me.
>
no filtering, the two artifacts most visible would be a 'crosshatch' dot pattern
in the luma and extreme amounts of chroma flashing as a result of luma detail.
I have played with this on my component D9 decks (but in component mode),
and the results aren't very good :-). For extremely low end situations, I
guess that the results might only be 'gross' and not totally
unacceptable. :-). (D9 can ALMOST handle a composite color signal on
the component Y-channel, but doesn't reproduce color totally correctly. Playback
on a Y-input B+W TV shows a dot pattern.)
Note that early B+W TVs (with very good video response) would show
a cross hatch dot pattern. This pattern is mitigated by a rolloff that
extinguishes most detail above 3MHz. The 'color flashing' is commonly
seen on non-comb filter TV sets, but it is possible that a RAW S-Video
input would even be worse.
The same 'crawling' dots that surround a comb filter TV's chroma areas
are extended into the full area of a color field.
John
Dan Harkless
news:<4189894b.0204...@posting.google.com>:
> Howdy. I'm building a new PC whose primary function [...] will be
> to DVD (i.e., 720x480 interlaced, 60 * 1000/1001 fields per second).
>
> Video sources will be TV programs from my Sony SAT-T60 DirecTiVo, my
> VHS video collection, and, sometime in the future, my LD collection
> (so I'll need composite and S-video in, but not component).
>
> My primary concern is quality -- for instance, I'll be buying a
> videophile or pro [S-]VHS deck with TBC so I can squeeze the maximum
> quality out of my VHS tapes and then be confident enough to throw away
> the originals.
On rec.video.desktop the main thread of my query above is being
carried out, but I wanted to include this group in a subquery about
the VCR.
My VHS collection consists of home movies, recordings from TV (using a
variety of VCRs, some mono, some Hi-Fi stereo) going back as far as 20
years ago (when we had a local TV store record a commercial for O.P.
Clothing I appeared in), some multigenerational "tape trader" dubs,
prerecorded movies and music video compilations, and a few odds and
ends.
As I mention above, I want to squeeze maximum possible quality out of
these tapes so I can get their content onto DVD-R or DVD+R and then
feel confident enough to throw away the originals. Because many of
these tapes are quite hoary or had low video quality to begin with,
and because the commercial tapes are Macrovision-protected, I'll need
a VCR with a TBC (or an external TBC, I suppose, but I've read that
with the possible exception of the ultra-high-end ones, those damage
picture content more than built-in ones).
*ALL* I care about in selecting this VCR are:
* Playback quality of VHS tapes.
* Ability to go through a few hundred tapes, a few on poor tape
stock, some as old as 20 years, without breaking down, and
hopefully without needing to be manually head-cleaned too often.
* A TBC that allows me to digitize Macrovision-protected material
and low-quality signals without dropping frames, and doesn't
significantly degrade the picture when engaged. Ability to use
the TBC in signal pass-through mode would be a nice bonus but is
not really required (my DVD player is modded).
Except in the case of all else being equal between two models, I don't
care about:
* Playback quality of S-VHS tapes (I don't have any).
* Record quality (or even ability to record).
* Tuner quality, editing ability, timecode support, or other
non-essential features.
I'd like to spend less than $1500. Here are all the
current-production pro and prosumer VCRs I'm familiar with that
potentially meet my requirements:
* JVC prosumer: HR-S9800U (list $699.95), HR-S9900U (list $599.95),
and HM-DH30000[U] (street $1500, but I'd be paying for D-VHS
abilities I may never have a use for).
* JVC pro: SR-V10U (list $400).
* Panasonic pro: AG-1980 (list $1599).
I'm not considering the JVC 7800 and 7900 since they only have a 2MB
frame buffer. I'm not considering the HR-DVS2U since I don't really
need another MiniDV device. While I'm not always adverse to used
equipment, used tape decks as a category give me the willies, so I'm
not considering any "eBay specials".
I would appreciate any help anyone can offer in choosing between the
above-listed models. Also, if there are any that I've missed (e.g.
because the manufacturer doesn't explicitly list "TBC" or "Time Base
Converter" in the specifications on their website), please let me
know. And if for some reason you think I'm making a mistake limiting
myself to decks with TBCs, and I should instead get the most expensive
external TBC I can afford, let me know that too.
Thanks in advance for any guidance!
--
Dan Harkless
use...@harkless.org
r.crowley
> Because many of these tapes are quite hoary or had low
> video quality to begin with, and because the commercial
> tapes are Macrovision-protected, I'll need a VCR with a
> TBC (or an external TBC, I suppose, but I've read that
> with the possible exception of the ultra-high-end ones,
> those damage picture content more than built-in ones).
If you are talking about the cheap consumer things then perhaps you're
right. But those aren't real TBCs anyway, they are just proc amps and
(maybe) sync regenerators.
Seems unlikely that you will find ANY VCR in your price range with a real
TBC. Besides, the machines with built-in TBCs likely are designed
specifically to EXCLUDE copy protection anomolies (i.e. Macrovision) for
obvious reasons.
For $1500, this is what I actually did:
1) Get a Panasonic 1980 (street $1000) and
2) Get a used professional TBC (I have bought a couple of Horitas for ~$500
on eBay)
Freddy Vegas
"r.crowley" <rcro...@xprt.net> wrote in message
news:ucchdg7...@corp.supernews.com...
John S. Dyson
> > Because many of these tapes are quite hoary or had low
> > video quality to begin with, and because the commercial
> > tapes are Macrovision-protected, I'll need a VCR with a
> > TBC (or an external TBC, I suppose, but I've read that
> > with the possible exception of the ultra-high-end ones,
> > those damage picture content more than built-in ones).
>
> If you are talking about the cheap consumer things then perhaps you're
> right. But those aren't real TBCs anyway, they are just proc amps and
> (maybe) sync regenerators.
>
> Seems unlikely that you will find ANY VCR in your price range with a real
> TBC. Besides, the machines with built-in TBCs likely are designed
> specifically to EXCLUDE copy protection anomolies (i.e. Macrovision) for
> obvious reasons.
>
than a time base corrector. It might not be 'perfect' like a true TBC, but the
JVC's (the 9600 for example) do write the video into a ram, and clock it
out correctly.
Consumer VCRS very often nowadays try to sense the Macrovision and other
copy protection flags. A true TBC (like my DPS) does clean everything up, but
my JVC 30000, JVC 9600 and DHR1000 all resist recording when there is
Macrovision present. In fact, even the SCC doesn't get rid of the stuff so that
the DHR1000 still detects copy protection. I don't know what the 30000 does
with the SCC (which is just a sync regen thing with a comb for correcting the
color.) I only played with the Macrovision issues when I first got some of my
equipment, so don't know about the 30000.
John
lhorwinkle
True, some of these will alter the video a bit. But, hey,
he's saying that his tapes are old and "hoary". So if the
tapes already suck, a little distortion from a Mac-eliminator
is no big deal.
"John S. Dyson" <dy...@iquest.net> wrote in message
news:Ttwx8.260$I04....@news1.iquest.net...
Dan Harkless
> > Because many of these tapes are quite hoary or had low
> > video quality to begin with, and because the commercial
> > tapes are Macrovision-protected, I'll need a VCR with a
> > TBC (or an external TBC, I suppose, but I've read that
> > with the possible exception of the ultra-high-end ones,
> > those damage picture content more than built-in ones).
>
> If you are talking about the cheap consumer things then perhaps you're
> right. But those aren't real TBCs anyway, they are just proc amps and
> (maybe) sync regenerators.
Well, I guess since I'm not trying to sync multiple tape decks
together, I don't need a "real" TBC (unless they're the only ones that
don't appreciably damage the signal).
> Seems unlikely that you will find ANY VCR in your price range with a real
> TBC. Besides, the machines with built-in TBCs likely are designed
> specifically to EXCLUDE copy protection anomolies (i.e. Macrovision) for
> obvious reasons.
Yeah, I guessed the consumer/prosumer ones would probably have
explicit Macrovision detection so would prevent you from cleaning it
with their TBCs, but I thought this wouldn't be an issue for the pro
decks (the way SCMS is defeatable on pro digital audio recorders but
not on consumer ones). I mean, surely pros occasionally get into a
situation where their only source for some footage that they're
authorized to use is on a Macrovision-protected tape?
> For $1500, this is what I actually did:
> 1) Get a Panasonic 1980 (street $1000) and
> 2) Get a used professional TBC (I have bought a couple of Horitas for ~$500
> on eBay)
Sounds reasonable. Since TBCs are solid-state, I certainly am not
inherently adverse to buying a used one (though I'd have no real way
of knowing if it's functioning up-to-spec). Unfortunately I have no
way to judge which TBCs do the least damage to the signal and what
reasonable used prices for them would be.
What kinds of things are you able to do with your Horitas that you
can't do with your 1980's builtin TBC?
--
Dan Harkless
use...@harkless.org
Dan Harkless
> The high end consumer JVCs have more of a time base stabilizer for playback
> than a time base corrector. It might not be 'perfect' like a true TBC, but
> the JVC's (the 9600 for example) do write the video into a ram, and clock it
> out correctly.
Well, actually, my main concern with the TBC is that it do a good job
of making crappy old VHS tapes cleanly digitizable. Ability to remove
Macrovision from commercial tapes so I can digitize them is secondary
-- most of the tapes in my commercial collection have been released on
LaserDisc (which doesn't have Macrovision protection) or DVD, meaning
I can just throw out the VHS (assuming I can find a used LaserDisc
copy, when necessary).
There will be some Macrovision-protected things that were never
released on any format besides VHS (particularly some of the music
video collections, I'm guessing), but I suppose if it's not possible
to get a TBC within my budget that can deal with those, I can always
rent a pro one to use on just those tapes.
> Consumer VCRS very often nowadays try to sense the Macrovision and other
> copy protection flags. A true TBC (like my DPS) does clean everything up,
> but my JVC 30000, JVC 9600 and DHR1000 all resist recording when there is
> Macrovision present. In fact, even the SCC doesn't get rid of the stuff so
> that the DHR1000 still detects copy protection. I don't know what the 30000
> does with the SCC (which is just a sync regen thing with a comb for
> correcting the color.) I only played with the Macrovision issues when I
> first got some of my equipment, so don't know about the 30000.
I don't care whether the SVHS deck I get refuses to record
Macrovision-protected material. I only care whether my digitizer
refuses to record it (or records it with Macrovision artifacts like
brightness sine waving) after the signal is passed through the SVHS's
deck's TBC (or an external one, if I have to go that route).
I haven't decided yet what my digitizer will be, but I've narrowed it
down to:
* ViewCast Osprey-2x0
* Sony DCR-PC5 MiniDV camcorder (already owned)
* Matrox RT2500
* Canopus DVStorm SE Plus
Anyone know whether any of these is particularly good or particular
bad at digesting Macrovision-protected material (or
poor-quality/multigenerational non-Macrovisioned VHS material)?
--
Dan Harkless
use...@harkless.org
John S. Dyson
>
> I don't care whether the SVHS deck I get refuses to record
> Macrovision-protected material. I only care whether my digitizer
> refuses to record it (or records it with Macrovision artifacts like
> brightness sine waving) after the signal is passed through the SVHS's
> deck's TBC (or an external one, if I have to go that route).
>
that can add some motion artifacting. This isn't exactly what you are
dealing with, but for recording 'other' 'composite' sources, I have found
that initially the 3D comb looks okay, but then the motion artifacts start
becoming obvious.
The issue of archiving 'old' material is one that I have been researching
for quite a while. I am staging my 'old' stuff onto D9 (essentially lossless)
after TBC'ing it. The next step will be non-realtime DVD production on
a PC. The technology associate with DVD allows for non-realtime noise
reduction and encoding that makes a full tradeoff to provide maximum
quality.
Old material will tempt you to enable noise reduction also. Sometimes
the noise reduction schemes also add motion artifacting.
What I am 'warning' is that once something like noise reduction is done
to old tapes or laser disks, or if a motion artifact prone 3D comb is applied
to one of your 'treasures', then that artifact will stay on the video for all
future generations.
The whole issue of keeping copies of some 'oldies but goodies', perhaps
slightly improving the original, but not adding more problems than the
original is full of non-trivial tradeoffs. I don't have any perfect answers
yet.
John
r.crowley
> can't do with your 1980's builtin TBC?
Mainly proc amp controls:
* Luminance Gain (consumerspeak: "contrast")
* Luminance Pedestal [DC offset] (consusmerspeak: "brightness")
* Chroma Gain (consumerspeak: "color")
* Chroma Phase (consumerspeak: "hue")
Certainly seems like you will need these to compensate for your
lower-quality source tapes!
And then, of course, Macrovision elimination(!)
Even an out-of-spec pro TBC is better than a new consumer toy "proc amp"!
I've bought four used pro TBCs on eBay and never had a problem with any of
them. Many sellers are able to test them or at least verify functionality
before you bid.
QL
rental tapes "protected"?
--
QL
"Dan Harkless" <use...@harkless.org> wrote in message
news:4189894b.02042...@posting.google.com...
r.crowley
> Will Premiere capture a macrovison protected tape?
It is a function of the hardware (i.e. you video capture equipment), not a
function of the software (Premiere) Generally, NO they will NOT capture
protected source material (DVD, tape, sattelite, cable, etc.)
It seems unlikely that ANY of your choices for equipment will capture
Macrovision-encoded source material.
> Are most/all of the rental tapes "protected"?
Likely 100% or nearly so, but I don't rent many tapes.
I've only tried capturing Macrovision once (and had to use a professional
TBC to eliminate the "encryption") No, I didn't make an illegal copy, I was
trying to capture a single frame image.
Ole Hansen
> * ViewCast Osprey-2x0
>
> * Sony DCR-PC5 MiniDV camcorder (already owned)
>
> * Matrox RT2500
>
> * Canopus DVStorm SE Plus
>
> Anyone know whether any of these is particularly good or particular
> bad at digesting Macrovision-protected material (or
> poor-quality/multigenerational non-Macrovisioned VHS material)?
The Canopus is said to be highly sensitive to Macrovision, even detecting
it sometimes when none is present (e.g. noisy tape).
The BT87x chips do *not* have Macrovision circuitry built in, nor do they
have AGC. Like an old VCR, therefore, they are probably fairly immune to
Macrovision.
Ole
Ole Hansen
> ... because the commercial tapes are Macrovision-protected, I'll need
> a VCR with a TBC ...
[...]
> * A TBC that allows me to digitize Macrovision-protected material
[...]
> (my DVD player is modded).
Uh, oh.
Paul Tauger, a copyright lawyer who hung around here a few weeks ago,
convinced us that, in the US, you cannot legally copy copyrighted *video*
material without permission, even if you purchased a copy and even if it
is for your own use only. In other words, "medium shifting" video (e.g.
from VHS -> DVD-R, even if the VHS is yours) is illegal. Sad, but true.
Audio material is an exception, i.e. cassette tape -> CD-R is ok if for
personal use.
Defeating Macrovision is another thing US lawmakers don't want you to do.
I think (not sure) that even discussing technical details about this is
illegal.
So, maybe you need/want/should rethink your plans .... ?!?
Disclaimer: I am not a lawyer, and this is not legal advice.
Anyway, just thought I'd bring it up for a livelier discussion ;)
Ole
Ole Hansen
> But I suppose that's a moot point if the Osprey *drivers* have all the
> advantages that you mention. Perhaps ViewCast is more of a software
> house than a hardware operation.
That is my impression, too.
> BTW, Tim Newton mentioned in his post in this thread that one
> shouldn't expect perfect audio sync from the Osprey-2x0, even though
> it presumably uses a master oscillator for both audio and video. Is
> this your experience as well,
Yes. Indeed, the audio and video oscillators on my Osprey 200 do not seem
to be perfectly synched. The audio clock is about 30ppm too fast
(i.e. runs at 44101.3 Hz vs. 44100 Hz). If I did the math correctly, this
results in one duplicated frame every 37 minutes (if the video clock is
29.97 Hz).
The card appears to have a separate on-board audio digitizer, a tiny
micro IC. It seems to use a highly precise clock, which is not
synchronized to the video clock, however.
For comparison, my ISA sound card (AWE64) is -1530ppm off!!! The Osprey is
a precision instrument compared to that.
> or is there a capture program one can
> use (e.g. iuVCR) with the board that does the correct 48 kHz 8008
> samples per 5 frames of 29.97 NTSC (see DV magazine's great article
> "Keeping That Syncing Feeling",
> <http://www.dv.com/features/features_item.jhtml?LookupId=/dv/xml/feature/2001/meyer0901>)
> and that also breaks the 2 GB and 4 GB filesize limits (unlike the
> Osprey's bundled capture software, per Tim)?
I use VirtualDub with the audio-sync patch. I posted the link before.
With that patch, audio sync headaches disappear completely and forever.
Additionally, VirtualDub can capture beyond 2/4GB without any problems.
BTW, the Osprey card can only sample audio at 44100 Hz, nothing else.
If you want native 48 kHz sampling, I would use a separate sound card.
Ole
Dan Harkless
> > BTW, Tim Newton mentioned in his post in this thread that one
> > shouldn't expect perfect audio sync from the Osprey-2x0, even though
> > it presumably uses a master oscillator for both audio and video. Is
> > this your experience as well,
>
> Yes. Indeed, the audio and video oscillators on my Osprey 200 do not seem
> to be perfectly synched. The audio clock is about 30ppm too fast
> (i.e. runs at 44101.3 Hz vs. 44100 Hz). If I did the math correctly, this
> results in one duplicated frame every 37 minutes (if the video clock is
> 29.97 Hz).
>
> The card appears to have a separate on-board audio digitizer, a tiny
> micro IC. It seems to use a highly precise clock, which is not
> synchronized to the video clock, however.
Gosh, well that's silly! What a wasted opportunity to eliminate the
pernicious audio sync issue.
> For comparison, my ISA sound card (AWE64) is -1530ppm off!!! The Osprey is
> a precision instrument compared to that.
Yeah, I suppose audio 1 frame out-of-sync with the video probably
isn't perceptible for most viewers. Hopefully VirtualDub Sync isn't
written in such a way that it's harder to correct a very tiny sync
slip than a larger one (e.g. how it's easier to scale an image by 1.5x
without artifacts than by 1.01x). Hopefully only the points at which
samples are dropped or added are mangled, and most of the PCM stream
is untouched.
> > or is there a capture program one can
> > use (e.g. iuVCR) with the board that does the correct 48 kHz 8008
> > samples per 5 frames of 29.97 NTSC (see DV magazine's great article
> > "Keeping That Syncing Feeling",
> > <http://www.dv.com/features/features_item.jhtml?LookupId=/dv/xml/feature/2001/meyer0901>)
> > and that also breaks the 2 GB and 4 GB filesize limits (unlike the
> > Osprey's bundled capture software, per Tim)?
>
> I use VirtualDub with the audio-sync patch. I posted the link before.
> With that patch, audio sync headaches disappear completely and forever.
> Additionally, VirtualDub can capture beyond 2/4GB without any problems.
Great to hear. I'm still a bit nervous, though, as to whether any of
the non-DV hardware/software combinations does the correct samples per
frame sequencing discussed in that article I referenced above. For 48
kHZ it's supposed to be:
Frame 1 Frame 2 Frame 3 Frame 4 Frame 5 Frame 6+
------- ------- ------- ------- ------- --------
1602 1601 1602 1601 1602 (repeat)
samples samples samples samples samples
> BTW, the Osprey card can only sample audio at 44100 Hz, nothing else.
> If you want native 48 kHz sampling, I would use a separate sound card.
Whoa!! Thanks for the warning! The supported sample rates aren't
mentioned on (the HTML part of) ViewCast's site! Only in the data
sheet:
http://www.viewcast.com/download/OS200datasheet.pdf
I just presumed the 200 would support 48 kHz since that's the standard
sampling rate for DV, DVD, etc. 44.1 isn't even supported for DVD, no
doubt for the reasons mentioned in that DV Magazine article:
Look at the ratio for 44.1kHz audio against (30,000 ÷ 1001) fps
NTSC color video: 147,147 samples per 100 frames. That's really
ugly, and requires a complex samples-per-frame sequence--which
translates to far more complicated electronics and software. This
is a key reason why 44.1kHz audio sample rates are not used for
professional video production. Some manuals for DV cameras refer
to the 44.1kHz rate as a consumer rate and don't promise sync
between 44.1kHz audio and video. If you want to be safe when
working with video, you really should work with 48kHz audio, as
many video editors already do.
It would suck to have to sample-rate-convert all my captures from 44.1
to 48 prior to DVD burning (and suffer the artifacts and potential
sync problems that would entail). Luckily all is not lost -- the
Osprey-210 and 220 offer 48 kHz sampling:
http://www.viewcast.com/download/OS210220datasheet.pdf
Looks like Osprey-210 is my card (if I indeed decide to go the
uncompressed route rather than a DV encoding solution). The 220 would
be nice if I get a pro VCR like the Panasonic AG-1980 with balanced
outputs, but it's $299 (direct from ViewCast) to the 210's $199 just
for that feature! There's no way balanced audio support costs
anywhere near $100 more -- they must just up their estimate of what
the market can bear with the 220 since they expect only pros will want
the balanced audio support.
--
Dan Harkless
use...@harkless.org
Dan Harkless
> Macrovision eliminators are available for $100 or so.
>
> True, some of these will alter the video a bit. But, hey,
> he's saying that his tapes are old and "hoary". So if the
> tapes already suck, a little distortion from a Mac-eliminator
> is no big deal.
No, I said some of them are. The really hoary ones are old recordings
from TV or "tape trader" tapes that don't have Macrovision. The
Macrovision-protected tapes are in as good a shape as can be expected,
and I don't want to lose any more quality than necessary (otherwise I
wouldn't bother to get a prosumer/pro VCR, try for uncompressed
capture rather than going through DV, etc.).
--
Dan Harkless
use...@harkless.org
Paul Tauger
"Ole Hansen" <o...@redvw.com> wrote in message
news:Pine.LNX.4.44.020425...@linus.redvw.com...
> On 23 Apr 2002, Dan Harkless wrote:
>
> > ... because the commercial tapes are Macrovision-protected, I'll need
> > a VCR with a TBC ...
>
> [...]
>
> > * A TBC that allows me to digitize Macrovision-protected material
>
> [...]
>
> > (my DVD player is modded).
>
> Uh, oh.
>
> Paul Tauger, a copyright lawyer who hung around here a few weeks ago,
Actually, I've been hanging around here for a couple of years, but that's
okay -- it's nice to be remembered. ;)
> convinced us that, in the US, you cannot legally copy copyrighted *video*
> material without permission, even if you purchased a copy and even if it
> is for your own use only.
That's not what I said at all. There are numerous circumstances under which
you can copy protected videos. Making duplicate copies of DVDs and
VHS tapes for medium-shifting isn't one of them, however.
>In other words, "medium shifting" video (e.g.
> from VHS -> DVD-R, even if the VHS is yours) is illegal. Sad, but true.
Perhaps Congress will change this -- they did with regard to audio tapes.
It is perfectly permissible to take an audio CD and "rip" MP3s or make tape
cassettes for your personal use.
>
> Audio material is an exception, i.e. cassette tape -> CD-R is ok if for
> personal use.
Ah, you're way ahead of me. ;)
>
> Defeating Macrovision is another thing US lawmakers don't want you to do.
> I think (not sure) that even discussing technical details about this is
> illegal.
That's a very good question. There is a tension between the Digital
Millenium Copyright Act, which makes illegal devices for defeating
encryption, and the Audio Home Recording Act, which defines as
"non-infringement," making audio copies.
You're also right -- in theory, even talking about defeating Macrovision
could, I suppose, result in a violation of the DMCA. However, there is an
established doctrine for defense to contributory infringement called
"substantial non-infringing use." This was a key provision of the famous
Sony Betamax case (Sony v. Universal).
It might work something like this. Macrovision-encoded material results in
a green color shift on my Mits RP television. I don't know why that should
be -- my guess is it has something to do with the way Macrovision messes
with synchronization (I'm not a video expert, so bear with me if my
explanation isn't quite right). There are a variety of devices that reshape
and/or regenerate sync signals. A time base corrector will do this, as will
Sima's color enhancer and a few other products. I find that these work
quite well for restoring the normal color balance to my Mits when playing
heavily-Macrovision encoded material.
Dan Harkless
> The Canopus is said to be highly sensitive to Macrovision, even detecting
> it sometimes when none is present (e.g. noisy tape).
Ouch.
> The BT87x chips do *not* have Macrovision circuitry built in, nor do they
> have AGC. Like an old VCR, therefore, they are probably fairly immune to
> Macrovision.
Great!! Well, that settles that, then. The Osprey-210 it is...
Thanks so much for all your input, Ole!
--
Dan Harkless
use...@harkless.org
Dan Harkless
> "Dan Harkless" <use...@harkless.org> wrote:
> > What kinds of things are you able to do with your Horitas that you
> > can't do with your 1980's builtin TBC?
>
> Mainly proc amp controls:
> * Luminance Gain (consumerspeak: "contrast")
> * Luminance Pedestal [DC offset] (consusmerspeak: "brightness")
> * Chroma Gain (consumerspeak: "color")
> * Chroma Phase (consumerspeak: "hue")
>
> Certainly seems like you will need these to compensate for your
> lower-quality source tapes!
Yes, those will need to get fixed at some stage, but it's my
understanding that capture cards like the Osprey-210 I'm now planning
on using allow adjustment of those same parameters. And even if they
didn't those could be fixed in software after capture (though of
course extreme adjustments to those in the digital domain could result
in lower quality than having done it in the analog domain prior to
capture).
> And then, of course, Macrovision elimination(!)
Okay. Thanks for the confirmation that the 1980's TBC doesn't handle
that.
> Even an out-of-spec pro TBC is better than a new consumer toy "proc amp"!
Right. But I wasn't considering any of the $100 "picture stabilizers"
as an alternative to a pro TBC, just the builtin TBCs in prosumer and
pro VCRs.
> I've bought four used pro TBCs on eBay and never had a problem with any of
> them. Many sellers are able to test them or at least verify functionality
> before you bid.
Unfortunately my personality and penchant for researching the hell out
of any major purchase precludes me from just buying any random pro TBC
I find on eBay as long the seller claims it's functional. I'd have to
do a little research to find out which pro TBCs are well-regarded and
which ones aren't, and what the market values are. And researching a
relatively niche-use pro piece of equipment like this where the
quality difference between models is tricky to quantify will not be
easy...
--
Dan Harkless
use...@harkless.org
Dan Harkless
> Uh, oh.
>
> Paul Tauger, a copyright lawyer who hung around here a few weeks ago,
> convinced us that, in the US, you cannot legally copy copyrighted *video*
> material without permission, even if you purchased a copy and even if it
> is for your own use only. In other words, "medium shifting" video (e.g.
> from VHS -> DVD-R, even if the VHS is yours) is illegal. Sad, but true.
Hmm. Thanks for the pointer. Here's Paul's main thread on this, in
case others besides me missed it the first time around:
http://groups.google.com/groups?threadm=9eccb616.0203211437.3ec4565a%40posting.google.com
> Audio material is an exception, i.e. cassette tape -> CD-R is ok if for
> personal use.
Okay. I didn't realize that there were explicit home copying rights
assigned to audio and computer programs that don't apply to other
media. I did go back and annotate a post I made on another (set of)
newsgroup(s), which was discussing the legality of making personal-use
copies of owned DVDs, with a pointer to the March "copying videos"
thread.
> Defeating Macrovision is another thing US lawmakers don't want you to do.
> I think (not sure) that even discussing technical details about this is
> illegal.
>
> So, maybe you need/want/should rethink your plans .... ?!?
>
> Disclaimer: I am not a lawyer, and this is not legal advice.
Well, I guess the "escape hatch" in this body of law is the fair use
doctrine. I take it from reading the "copying videos" thread that
there is no case law establishing whether medium shifting in the video
domain comes under fair use or not. As stated in that thread, the
likelihood of me personally being sued for medium shifting seems
exceedingly remote. And even if I were sued, the optimist in me would
hope that something that so clearly falls under "fair use" from the
layman's perspective would also be legally considered as such.
In any case, as I stated, the only commercial VHS tapes I intend to
put on DVD[-+]R are ones for which there is no commercial DVD version
available. Most or all of them (if there turn out to BE any, to begin
with) will end up being titles which are out-of-print in ALL formats,
including VHS. As far as I know, that's not a license to copy, under
the law, but I'm sure it'd be a positive factor in showing fair use.
--
Dan Harkless
use...@harkless.org
Dan Harkless
[...]
> Howdy. I'm building a new PC whose primary function (besides the
> usual DTP, web browsing, etc.) will be uncompressed /
> non-lossy-compressed capture of analog NTSC video for burning to DVD
> (i.e., 720x480 interlaced, 60 * 1000/1001 fields per second).
Man, it really annoys me that Google Usenet breaks threads up if you
change the subject line. Why don't they obey the References: line??
That's part of the reason it exists -- so you can modify the subject
in a subthread to be more descriptive, without breaking threading.
In any case, if anyone was following my original thread but missed the
"Best < $1500 TBC-equipped VCR for digitization of VHS collection?"
subthread, it's here:
http://groups.google.com/groups?threadm=4189894b.0204232123.7443f62c%40posting.google.com
I also posted a parallel version of that subthread in alt.video.vcr
and alt.home-theater.misc (didn't just crosspost onto them from the
main subthread since not all sites get all the alt.* groups and
following up to a post that's crossposted to locally non-existent
newsgroups fails on many/most newsreaders):
http://groups.google.com/groups?threadm=4189894b.0204232123.7443f62c%40posting.google.com
Annoyingly, Google Usenet's "we have omitted some entries very similar
to" feature hides all postings from the parallel thread if you do a
search, even though they are technically distinct threads.
Anyway, just wanted to help out anyone now or in the future that's in
my same boat and wanted to read ALL the information in this thread and
its subthreads.
--
Dan Harkless
use...@harkless.org
Dan Harkless
digitization, I was really only referring to my TV recordings. The
commercial material I will keep the originals of, so I can prove that
I'm a legitimate owner, if it were ever to come up.
Of course, one might imagine that documenting myself destroying the
originals would also suffice. Then again, they could claim I used the
desktop video production software I've talked about buying on this
group to fake that footage. ;^>
--
Dan Harkless
use...@harkless.org
Dan Harkless
> that can add some motion artifacting.
John, I did some searching and I couldn't find any discussion of
motion artifacts introduced by comb filters. What do they look like,
and how obtrusive are they? I presume a pro deck like the Panasonic
AG-1980 would have a higher-quality comb filter than the prosumer
decks and thus less artifacting?
> What I am 'warning' is that once something like noise reduction is done
> to old tapes or laser disks, or if a motion artifact prone 3D comb is applied
> to one of your 'treasures', then that artifact will stay on the video for all
> future generations.
Okay. Your warning is well-taken (at least to the extent that that's
possible considering I have no idea what these artifacts are that
you're warning against ;^>).
> The whole issue of keeping copies of some 'oldies but goodies', perhaps
> slightly improving the original, but not adding more problems than the
> original is full of non-trivial tradeoffs.
No kidding. I'm used to doing exhaustive research, but researching
this purchase (and process) has been exhaustING as well.
--
Dan Harkless
use...@harkless.org
John S. Dyson
> "John S. Dyson" <dy...@iquest.net> wrote:
> > One comment here also: some SVHS decks have 3D comb filters
> > that can add some motion artifacting.
>
> John, I did some searching and I couldn't find any discussion of
> motion artifacts introduced by comb filters. What do they look like,
> and how obtrusive are they?
>
perceived most often is motion smear. 3D comb filters do average over
2 fields or 2 frames (or more), and this does cause smear. This is one
(of many) reasons why 50Hz 3D combs are trickier than 60Hz 3D
combs.
>
>I presume a pro deck like the Panasonic
> AG-1980 would have a higher-quality comb filter than the prosumer
> decks and thus less artifacting?
>
I don't know if they are that much better. My D9 decks have GOOD
2D combs that allow for reasonable tearing apart and reasonable
reassembly of NTSC composite. I prefer the 2D comb behavior on
the D9 decks and the 3D comb on the HM30000 over and above the
3D comb on the 9600. It appears (from casual observation) that the
CBS spots on HDTV use damned good 2D combs rather than 3D combs.
It is possible that if they use 3D combs, they have supurb motion
compensation. (The D9 deck 2D combs are good enough that I can
record composite, subsequently playing back the NTSC composite
without compromising the 3D comb on my HDTV. Lower end 2D
combs or laggy 3D combs cause numerous artifacts. If playing back
a laserdisk recorded by a recorder with a 2D comb, even a 3D comb
on the TV cannot correct the degradation from the low end 2D comb in the
path.)
Damned good 2D combs will still have color flashing, but it will be
restricted to certain kinds of diagonal spatial entities. If the recording
device can adequately reassemble the video and consider the
damage done by a 2D comb, then the composite output of the recording
device will be VERY CLOSE to the original. It seems like alot of high-end
comb filters (not the SUPER HIGH END), use innovative 2D combs, probably
because of the complexity of dealing with the motion issues on 3D combs.
> > What I am 'warning' is that once something like noise reduction is done
> > to old tapes or laser disks, or if a motion artifact prone 3D comb is applied
> > to one of your 'treasures', then that artifact will stay on the video for all
> > future generations.
>
> Okay. Your warning is well-taken (at least to the extent that that's
> possible considering I have no idea what these artifacts are that
> you're warning against ;^>).
>
Basically, it looks like motion smear. Sometimes the first generation
of a 3D comb looks okay when being played back on a 2D comb TV,
but if played on a 3D comb TV, the motion smear builds stronger than
it should be. The action of a 3D comb is only operative on NTSC
composite, but video noise reducers can impose similar artifacts to
3D combs. I was suckered in initially by the 3D comb on my JVC9600,
and ended up with D9 recordings (I used the 3D comb from the SVHS
deck for recording composite) that had the motion smear. If I had
used the built-in D9 comb, the composite output from the D9 deck
would have been nearly identical to the original, and then the comb
would have been deferred to the final print.
One other 3D comb artifact that is more obvious than the smear would
be slightly unoptimal reversion to 2D comb mode. The 3D comb removes
the color flashing very well, but if there is movement in the video, then
the 2D comb artifacts become operative. I suspect that some 3D combs
don't have quite the sophistication in 2D comb mode that they should
have.
John
r.crowley
> Looks like Osprey-210 is my card (if I indeed decide to go the
> uncompressed route rather than a DV encoding solution). The 220 would
> be nice if I get a pro VCR like the Panasonic AG-1980 with balanced
> outputs,
All the Panasonic AG-19*0s I've ever seen (including my three) have
consumer-type RCA unbalanced audio in and out.
"Not that there's anything WRONG with that!" :-)
> but it's $299 (direct from ViewCast) to the 210's $199 just
> for that feature! There's no way balanced audio support costs
> anywhere near $100 more -- they must just up their estimate of what
> the market can bear with the 220 since they expect only pros will want
> the balanced audio support.
Sounds right to me!
r.crowley
> Unfortunately my personality and penchant for researching the hell out
> of any major purchase precludes me from just buying any random pro TBC
> I find on eBay as long the seller claims it's functional. I'd have to
> do a little research to find out which pro TBCs are well-regarded and
> which ones aren't, and what the market values are. And researching a
> relatively niche-use pro piece of equipment like this where the
> quality difference between models is tricky to quantify will not be
> easy...
OK, but if you think that the Panasonic AG-1980 is "pro" (as you said in two
current posts), your concept of what constitutes "pro",
"prosumer/industrial", and "consumer" doesn't correlate very well with the
mainstream concept. Not knocking the 1980, I own three of them, but nobody
I know thinks they are "pro". Your initial $1500 budget put you at the
BOTTOM end of the "prosumer" range, well removed from anything that could be
considered "pro".
My gut feeling is that the 4-line TBC in the 1980 will be insufficient to
recover/capture your older and grungier source tapes. But take your
worst-case tape and try it on some combinations of equipment to see how it
will work. Nobody can argue with real-world success.
Dan Harkless
> > John, I did some searching and I couldn't find any discussion of
> > motion artifacts introduced by comb filters. What do they look like,
> > and how obtrusive are they?
>
> They either look like motion smear or multiple images. What I have
> perceived most often is motion smear. 3D comb filters do average over
> 2 fields or 2 frames (or more), and this does cause smear. This is one
> (of many) reasons why 50Hz 3D combs are trickier than 60Hz 3D
> combs.
I see. And when you say "multiple images", you mean similar to
ghosting, or...?
> Damned good 2D combs will still have color flashing, but it will be
> restricted to certain kinds of diagonal spatial entities.
"Color flashing"?
> Sometimes the first generation
> of a 3D comb looks okay when being played back on a 2D comb TV,
> but if played on a 3D comb TV, the motion smear builds stronger than
> it should be. The action of a 3D comb is only operative on NTSC
> composite, but video noise reducers can impose similar artifacts to
> 3D combs.
Okay. Well, since the output of the SVHS deck will be burned to DVD,
which is inherently a component format, and I don't expect to ever
hook up a DVD player using its composite output, I'm not too worried
about stacked comb filter effects, just the quality of the 1st-gen.
comb output.
--
Dan Harkless
use...@harkless.org
HighPeaksVideo
DPS (Digital Processing Systems) made the personal tbc series (ISA CARD) from
one thru four. All have S INPUT, only the four has S output. I think you can
still buy the tbc4 card for about $850.00 new.
I use mine in an Amiga computer. DPS made (make ?) a rack mount controller box
for two cards also.
Craig H.
Dan Harkless
> All the Panasonic AG-19*0s I've ever seen (including my three) have
> consumer-type RCA unbalanced audio in and out.
Ah! Thanks for the correction. Panasonic doesn't specify what kind
of audio outs they have on their website:
http://www.panasonic.com/PBDS/subcat/Products/vtrs_vcrs/f_ag-1980.html
so I guess I was just assuming based on looking at the specs for other
pro decks like the (out of my price range) Sonys:
http://bpgprod.sel.sony.com/bpcnav/app/99999/16/118.99999.subcat.BPC.html
Cool. Now I won't feel a twinge for buying the Osprey-210 rather than
the 220.
--
Dan Harkless
use...@harkless.org
John S. Dyson
> "John S. Dyson" <dy...@iquest.net> wrote:
> > > John, I did some searching and I couldn't find any discussion of
> > > motion artifacts introduced by comb filters. What do they look like,
> > > and how obtrusive are they?
> >
> > They either look like motion smear or multiple images. What I have
> > perceived most often is motion smear. 3D comb filters do average over
> > 2 fields or 2 frames (or more), and this does cause smear. This is one
> > (of many) reasons why 50Hz 3D combs are trickier than 60Hz 3D
> > combs.
>
> I see. And when you say "multiple images", you mean similar to
> ghosting, or...?
>
a dynamic ghosting with movement.)
> > Damned good 2D combs will still have color flashing, but it will be
> > restricted to certain kinds of diagonal spatial entities.
>
> "Color flashing"?
>
If you take a look at a non-comb color TV, and note the color flashes in
areas of large black/white detail (e.g. US football referees), then that is
the kind of flashing that I speak of. It is actually the contamination of
the chroma by luma detail. Good 2d combs can hide the flashing in most
circumstances, while not incurring the lag from a low end 3d comb. The
best situation is where there is a full quality 2d comb in conjunction with
a good dynamic 3d combing mechanism. It seems like in many cases,
the quality of the 2d performance is compromised because the 3d comb
does hide alot of evils. If the 2d performance was maxed out, then the
3d comb's sensitivity to motion could be improved, therefore helping
to mitigate lag effects.
> > Sometimes the first generation
> > of a 3D comb looks okay when being played back on a 2D comb TV,
> > but if played on a 3D comb TV, the motion smear builds stronger than
> > it should be. The action of a 3D comb is only operative on NTSC
> > composite, but video noise reducers can impose similar artifacts to
> > 3D combs.
>
> Okay. Well, since the output of the SVHS deck will be burned to DVD,
> which is inherently a component format, and I don't expect to ever
> hook up a DVD player using its composite output, I'm not too worried
> about stacked comb filter effects, just the quality of the 1st-gen.
> comb output.
>
Yes -- I understand that in this specific instance. The 3d comb results
even at the first generation can be noticeable, but the benefits might
make a good tradeoff. IMO, it is important to be VERY CRITICAL of
the various schemes, so as to help spur on the manufacturers to
improve quality.
John
Dan Harkless
> I use a consumer JVC Svhs deck with DPS TBCs.
> DPS (Digital Processing Systems) made the personal tbc series (ISA CARD) from
> one thru four. All have S INPUT, only the four has S output. I think you can
> still buy the tbc4 card for about $850.00 new.
The motherboards I've been considering so far are without ISA slots,
but thanks for the suggestion.
--
Dan Harkless
use...@harkless.org
Dan Harkless
> > I see. And when you say "multiple images", you mean similar to
> > ghosting, or...?
>
> Depending upon the device, it can manifest as smear or multiple images (like
> a dynamic ghosting with movement.)
Wow, moving ghosts? That would be distracting...
> > "Color flashing"?
> >
> If you take a look at a non-comb color TV, and note the color flashes in
> areas of large black/white detail (e.g. US football referees), then that is
> the kind of flashing that I speak of.
Ah. Okay. I can picture that.
> IMO, it is important to be VERY CRITICAL of the various schemes,
> so as to help spur on the manufacturers to improve quality.
Thanks for enabling me to do so by making me more aware of these
issues with your posts.
--
Dan Harkless
use...@harkless.org
Dan Harkless
> OK, but if you think that the Panasonic AG-1980 is "pro" (as you said in two
> current posts), your concept of what constitutes "pro",
> "prosumer/industrial", and "consumer" doesn't correlate very well with the
> mainstream concept. Not knocking the 1980, I own three of them, but nobody
> I know thinks they are "pro". Your initial $1500 budget put you at the
> BOTTOM end of the "prosumer" range, well removed from anything that could be
> considered "pro".
Okay. I was just going by what the manufacturer considered them
(whether a particular model was in the consumer or pro areas of the
manufacturer's website). The AG-1980 is in Panasonic's pro area.
> My gut feeling is that the 4-line TBC
Ah, so it doesn't have a full frame-buffer, eh?
> in the 1980 will be insufficient to recover/capture your older and grungier
> source tapes.
So would you suggest buying a significantly cheaper source deck and
spending the money saved on the best external TBC I can get, or...?
> But take your worst-case tape and try it on some combinations of equipment to
> see how it will work. Nobody can argue with real-world success.
Well, that's easier said than done. I don't know that I have access
to any pro equipment to play with, short of actually buying something
(which I'd probably have to do via mail-order).
--
Dan Harkless
use...@harkless.org
Tim Newton
> uncompressed route rather than a DV encoding solution). The 220 would
> be nice if I get a pro VCR like the Panasonic AG-1980 with balanced
> outputs, but it's $299 (direct from ViewCast) to the 210's $199 just
What exactly is "balanced audio support"? I have the Osprey-220, and
I can help you a bit. The card does not sync audio perfectly with video
when capturing with VirtualDub. The software capture program they
supply does not support capture files beyond 2 gbytes, so it's basically
useless. I can capture with good audio sync with a commercial program
called AVI_IO, but AVI_IO crashes whenever there are lots of frame drops
and the resulting AVI file will not be readable after crashes.
The osprey-220 card will often swap fields, or miss a field, I don't
know, during a capture. If I review captured video in an editor,
I see that during a scene change there will be one frame from the old
scene, and one frame from the new scene. This is not due to telecining.
It occurs somewhat randomly; if I capture the same scene repeatedly,
sometimes it swaps the fiels during the scene change, sometimes not.
The good news is that if you use the older capture drivers, the problem
appears to go away.
Finally, there is a brightness/contrast problem that I could not solve
by merely adjusting the controls. It occurs when you capture a very
bright scene. For example, in a hockey game, most of the screen will
be bright white because of the ice, and the players will be dark.
These kinds of captures look like crap. I think the problem is that
the levels are too bright in the original source, and that screws up
the capture. Anyway, I ordered a color level corrector, which should
hopefully arrive today, so I can give that a try.
Overall, the Osprey-220 is better than the ATI AIW card, the Matrox
e450 TV, and the ADS Instant DVD; the only other capture cards that
I tested. Still it's not perfect.
So, I hope this helps. What exactly is "balanced audio support"?
Tim
Dan Harkless
> What exactly is "balanced audio support"? I have the Osprey-220, and
> I can help you a bit.
The easy answer: balanced audio means using the XLR connectors on your
220's breakout cable rather than the RCA audio connectors. Balanced
audio is better than unbalanced audio at bucking RF noise and ground
loop hum, going through long cable runs without issues, etc. Here's a
good rec.video.production thread on the use of balanced audio in the
video realm:
http://groups.google.com/groups?threadm=8p4m6g%24o3s%40portal.gmu.edu
> The card does not sync audio perfectly with video when capturing with
> VirtualDub.
Have you tried the improved VirtualDub Sync version
(<http://www-user.rhrk.uni-kl.de/~dittrich/sync/>) that Ole Hansen
reported getting problem-free results with?
> The software capture program they supply does not support
> capture files beyond 2 gbytes, so it's basically useless.
Pretty lame.
> The osprey-220 card will often swap fields, or miss a field, I don't
> know, during a capture.
Yuck. Ole was saying that his Osprey-200 _doesn't_ have the
dropped/reversed field problem that, for instance, Hauppage's drivers
have.
> The good news is that if you use the older capture drivers, the problem
> appears to go away.
Ah! That _is_ good news. Can you be more specific? Which versions
are okay? Are they still available for download from ViewCast? Do
they have other problems that the newer drivers are meant to fix?
> Finally, there is a brightness/contrast problem that I could not solve
> by merely adjusting the controls. It occurs when you capture a very
> bright scene. For example, in a hockey game, most of the screen will
> be bright white because of the ice, and the players will be dark.
> These kinds of captures look like crap. I think the problem is that
> the levels are too bright in the original source, and that screws up
> the capture.
And I assume you aren't using some lossy codec when you experience
this? I know some codecs (like the Microsoft DV codec) have
compressed/clipped brightness range issues like this.
> Anyway, I ordered a color level corrector, which should
> hopefully arrive today, so I can give that a try.
Assuming your problem is not due to a codec, I guess it suggests I
should indeed try to buy a proper pro TBC with level controls rather
than trying to make do with the builtin TBC of high-end SVHS deck.
> Overall, the Osprey-220 is better than the ATI AIW card, the Matrox
> e450 TV, and the ADS Instant DVD; the only other capture cards that
> I tested.
Good to know.
> Still it's not perfect.
Well, if I can't get it to work to my satisfaction, I guess I'll give
up on the uncompressed capture approach and just use my Sony DCR-PC5
to DV-encode my analog (or look at a Matrox RT2500 or Canopus DVStorm
SE).
--
Dan Harkless
use...@harkless.org
Ole Hansen
> The osprey-220 card will often swap fields, or miss a field, I don't
> know, during a capture. If I review captured video in an editor,
> I see that during a scene change there will be one frame from the old
> scene, and one frame from the new scene. This is not due to telecining.
> It occurs somewhat randomly; if I capture the same scene repeatedly,
> sometimes it swaps the fiels during the scene change, sometimes not.
I am not completely sure I understand. Do you mean that at a scene change
there is a frame that is a *mix* of old and new scenes?
What is the source (NTSC/PAL, interlaced/progressive)?
What do you mean by "swap fields"? Really: exchange even and odd lines?
Or: reverse the field order? I.e., does motion look smooth before the
suspect transition, then very jerky afterwards?
The drivers don't seem to enforce a fixed field order, so if you capture
repeatedly, you may get clips with different order. That could explain
what you see.
> Finally, there is a brightness/contrast problem that I could not solve
> by merely adjusting the controls. It occurs when you capture a very
> bright scene. For example, in a hockey game, most of the screen will
> be bright white because of the ice, and the players will be dark.
> These kinds of captures look like crap. I think the problem is that
> the levels are too bright in the original source, and that screws up
> the capture.
Several people have reported this problem, but somehow I can't reproduce
it with any of the three Bt878 cards I own. I can think of only two
explanations: indeed, your source is out of spec; or you have a defective
chip (or driver). Which input are you using? I always use S-video.
I wanted to mention one potential problem I *did* notice with my Bt878
cards. At scene transitions, sometimes there is some "bleeding" of the
previous scene into the next (or, of the next into the previous)!
Usually, the very first frame of the new scene "carries over" some color
or brightness from the previous frame. For instance, if there is a bright
light spot in one scene, then after the transition the first new frame may
still show a faint image of that spot. Often, only the top of the frame is
affected, and the effect gradually disappears along the vertical
direction. For instance, a transition from a dark room to a bright outdoor
scene may give a darkening or discoloring of the sky in the first new
frame. In terms of a blend ratio, I would guess that there is no more than
about 20% of a previous frame blended into the next.
This does not occur at every scene change, nor does it occur at every
dramatic scene change (e.g. dark -> bright, blue -> red, etc.) It also
does not seem to occur *during* a scene, in particular, there is no "tail"
following moving objects. The effect is indeed "somewhat random".
Initially, I suspected my VCR to be the cause, but further tests have
shown this to occur with all video sources and Bt cards.
One possibility that I haven't ruled out is a bug in Huffyuv 2.1.1. All
affected captures were done with that codec.
In any case, the effect is minor. It is extremely difficult to see these
"washed-out" transitions during playback.
Ole
Dan Harkless
> I wanted to mention one potential problem I *did* notice with my Bt878
> cards. At scene transitions, sometimes there is some "bleeding" of the
> previous scene into the next (or, of the next into the previous)!
> Usually, the very first frame of the new scene "carries over" some color
> or brightness from the previous frame. For instance, if there is a bright
> light spot in one scene, then after the transition the first new frame may
> still show a faint image of that spot. Often, only the top of the frame is
> affected, and the effect gradually disappears along the vertical
> direction. For instance, a transition from a dark room to a bright outdoor
> scene may give a darkening or discoloring of the sky in the first new
> frame. In terms of a blend ratio, I would guess that there is no more than
> about 20% of a previous frame blended into the next.
>
> This does not occur at every scene change, nor does it occur at every
> dramatic scene change (e.g. dark -> bright, blue -> red, etc.) It also
> does not seem to occur *during* a scene, in particular, there is no "tail"
> following moving objects. The effect is indeed "somewhat random".
>
> Initially, I suspected my VCR to be the cause, but further tests have
> shown this to occur with all video sources and Bt cards.
Certainly sounds like a 3-D comb filter artifact of the type John
Dyson has been talking about as opposed to an artifact of the
digitizer. I mean, the Bt8xx digitizes the signal serially, not in
parallel, right? It's not as if it's a CCD, so I don't see how it can
"remember" what was on a field at a given point by the time it gets
back around to that point on the next field.
Are you sure it even happens for non-comb-filtered Y/C input (e.g.
from a non-Macrovisioned DVD)?
--
Dan Harkless
use...@harkless.org
Tim Newton
> there is a frame that is a *mix* of old and new scenes?
Yes. Exactly. One field from the old scene, and one field from the
new scene.
> What is the source (NTSC/PAL, interlaced/progressive)?
Doesn't matter. I have noticed this effect capturing from a DVD, from
a commercially produced VHS tape, and capturing television programs from
the cable, all NTSC. I have captured the same scene change repeatedly,
and the problem has occurred on some captures, not on others.
> What do you mean by "swap fields"? Really: exchange even and odd lines?
> Or: reverse the field order? I.e., does motion look smooth before the
> suspect transition, then very jerky afterwards?
By swapping fields I meant whatever is causing the above to occur.
> Several people have reported this problem, but somehow I can't reproduce
> it with any of the three Bt878 cards I own. I can think of only two
> explanations: indeed, your source is out of spec; or you have a defective
> chip (or driver). Which input are you using? I always use S-video.
The same occurs with the S-video and the composite input. The problem
may be caused by out-of-spec sources. I recently bought a color corrector
and level adjuster, which I put into the line between the VCR and the
capture card, and it fixes the problem.
> I wanted to mention one potential problem I *did* notice with my Bt878
> cards. At scene transitions, sometimes there is some "bleeding" of the
> previous scene into the next (or, of the next into the previous)!
I noticed that too. I had assumed it was an artifcat generated by the
TBC. But, it didn't really bother me so I never looked into it any
further.
Also, on another note, often time when I capture with the Osprey-220,
one field is all 0xff00ff00. This occurs most often when capturing
from old EP-mode VHS tapes where the sync signals are not too good.
This is very noticeable during playback: on my TV screen, you see a
green "flash" everytime it occurs. I recently wrote a VirtualDub
filter to solve this, somewhat. The field data appears as 0xff00ff00,
repeatedly. So, whenever my filter sees that, it simply replaces that
field with the field from the previous frame. It's not perfect, but
it's a lot better.
Tim